Neurological Sciences

, Volume 34, Issue 3, pp 269–279

CCSVI and MS: no meaning, no fact


  • Claudio Baracchini
    • Department of Neuroscience-SSPNRRStroke Unit, First Neurology Clinic, University of Padova
    • Department of Neuroscience-SSPNRRMultiple Sclerosis Centre of the Veneto Region, First Neurology Clinic, University of Padova
  • Paolo Gallo
    • Department of Neuroscience-SSPNRRMultiple Sclerosis Centre of the Veneto Region, First Neurology Clinic, University of Padova
Review Article

DOI: 10.1007/s10072-012-1101-2

Cite this article as:
Baracchini, C., Atzori, M. & Gallo, P. Neurol Sci (2013) 34: 269. doi:10.1007/s10072-012-1101-2


A condition called “chronic cerebrospinal venous insufficiency” (CCSVI) has been postulated to play a role in the pathogenesis of multiple sclerosis (MS). This hypothesis implies that a complex pattern of extracranial venous stenosis determines a venous reflux into the brain of MS patients, followed by increased intravenous pressure, blood–brain barrier breakdown and iron deposition into the brain parenchyma, thus triggering a local inflammatory response. In this review, we critically analyze the scientific basis of CCSVI, the current literature on the relationship between CCSVI and MS, as well as the ultrasound methodology that has been claimed to provide evidence of impaired cerebral venous drainage. We show that no piece of the CCSVI theory has a solid supportive scientific evidence. The CCSVI appears to be a rather alien condition and its existence should be definitely questioned. Finally, no proven (i.e., based on strict scientific methodology and on the rules of evidence-based medicine) therapeutic effect of the “liberation” procedure (unblocking the extracranial venous obstruction using angioplasty) has been shown up to date.


CCSVIMultiple sclerosis


A disorder of cerebrospinal blood flow named “chronic cerebrospinal venous insufficiency” (CCSVI) has been proposed by Zamboni to play a possible etiopathogenetic role in multiple sclerosis (MS) [1]. This “big idea”, a term used by Zamboni himself to define his theory, arises from observations on systemic venous diseases, and the possible parallels between these and brain inflammation [2]. Zamboni’s working hypothesis is that brain inflammation is iron-dependent [3]; he postulated that multiple extracranial venous anomalies cause a venous reflux into the cerebrospinal compartment, determining an increased intravenous pressure that breaks down the blood–brain barrier (BBB), thus causing the deposition of iron in brain tissue and evoking a local inflammatory response. By applying five parameters of abnormal venous outflow, indicative of CCSVI, Zamboni et al. were able to demonstrate a strong relationship between CCSVI and MS. Indeed, in their pivotal study, they analyzed 109 patients with clinically definite MS and 177 control subjects by means of transcranial and extracranial color Doppler sonography and found that all patients with MS had abnormal venous parameters: the presence of at least two of those five parameters was observed as being diagnostic of MS with 100 % specificity, 100 % sensitivity, and positive and negative predictive values for MS of 100 %. Zamboni et al. went on to perform selective venography in 65 patients with MS as well as in some control subjects, and reported that patients with MS had multiple severe extracranial stenosis, while these abnormalities were never found in normal controls [4]. Furthermore, in a retrospective study, the same authors found that the distribution of the pathological hemodynamic patterns was highly predictive of the symptoms at onset and of the following clinical course [5].

By December 2011, several independent investigators have tried to reproduce—with various methodological approaches—the striking results obtained by Zamboni, but none have succeeded [619] However, despite his contradictory statements on the significance of CCSVI (that meanwhile has become “the chronic cerebrospinal venous insufficiency syndrome”) [20] and on the necessity of “liberating” MS patients, Zamboni et al. continue to advertise and publicize the CCSVI theory among MS patients. Often, the debate loses every scientific content, and becomes a simple and sophistic attempt to influence and condition the public and the “political” perception of CCSVI [21]. Indeed, scientists publishing “non-confirmatory studies” are pilloried (in the press, internet blogs, social network or in the web sites of “ad-hoc” pro-CCSVI associations) as corrupted by drug-producing industries or (even worse) as unqualified researchers, incapable of studying the cerebrospinal venous system.

In this review, we try to analyze critically the various aspects of Zamboni’s theory and address several questions not only on the relationship between CCSVI and MS, but also on the scientific basis of CCSVI and, thus, on its real existence. Of course, the real scientific platform is not based on “believe or not believe” but on “demonstrate or not demonstrate”.

The origin(s) of the theory

Preclinical data: Dr. T. J. Putnam’s dogs

In putting together the puzzle of his theory, Zamboni finds supporting evidence for CCSVI in remote preclinical and clinical studies performed by Putnam [2226] as well as in histological studies showing that the earliest evidence of microvascular injury in acute MS is the presence of focal endothelial cell-associated fibrin deposition [2729]. Putnam was convinced that the etiology of MS laid in the venous system and tested this hypothesis by tying the jugular veins of dogs. These animals developed a diffuse thrombosis of the cerebrospinal veins and a number of brain abnormalities, many of them “similar to encephalitis or MS”, and concluded that venular obstruction was the essential immediate antecedent to the formation of typical sclerotic plaques of MS. According to Putnam, thrombosis of the cerebral veins was also a common finding in brain tissue specimens from MS patients, who also often had venous thrombosis in the lung, in the liver, and in other organs. Of course, the knowledge on inflammatory brain diseases was very limited and immunology did not exist at Putnam’s time. Indeed, he defined “inflammation” as “a morbid condition consisting in congestion of blood vessels and exudation of plasma and blood corpuscles at the site of an infection or injury” [24]. According to Putnam, MS was included among the so-called “disseminated encephalomyelitis”, i.e., brain diseases whose lesions: “…(1) predominate in the white matter, (2) occur around engorged veins, (3) are characterized chiefly by myelin destruction and glial proliferations, but lymphocytic and hemorrhagic diapedesis may occur, (4) thrombosis of vessels is constantly observed in the acute stage and is presumably the origin of the changes in the parenchyma”. The firm conviction that “the primary abnormality was in the clotting mechanism of the blood” induced Putnam to suggest and try the use of anticoagulants in MS, and believed in this therapy until his death. Unfortunately, the disease defined as “disseminated encephalomyelitis” constituted a quite heterogeneous group of inflammatory disorders, including—besides MS—acute disseminated encephalomyelitis (post-infectious/post-vaccination), fulminant Marburg’s disease, optic neuromyelitis, Schilder’s disease, etc. Moreover, the clinical and instrumental criteria to evaluate the clinical outcome of MS patients were far to be available. Subsequent analysis of Putnam’s data showed no evidence of benefit from anticoagulant therapy in MS [30]. Other studies [31, 32] also gave negative results.

Of course, the dog model of MS proposed by Putnam is far from having a clear relationship with MS. Moreover, it also does not fit to the CCSVI theory, and it sounds astonishing that Putnam’s observations are used nowadays to support the CCSVI-based etiology of MS. Certainly, Putnam’s theory was the result of the medical knowledge in the 30s and the old concept (still dominant at that time) that manipulation of the blood could be useful in the treatment of many diseases including those affecting the nervous system. Before Zamboni’s big idea, however, the possible etiopathogenetic relevance of cerebrospinal venous reflux in MS was hypothesized by Schelling [33] that, however, did not furnish personal scientific observations supporting this hypothesis.

While this manuscript was under review, the results of an experimental study on mice whose jugular veins were ligated bilaterally (JVL group) and then observed for up to 6 months have been published online [34]. The mice were evaluated using CT venography and (99 m) Tc-exametazime to assess for structural and hemodynamic changes, and to evaluate signs of BBB breakdown and neuroinflammation. Moreover, flow cytometry and histopathology were performed to assess inflammatory cell populations and demyelination. No evidence of demyelination was found, and the mice remained without clinical signs. Despite the structural and hemodynamic changes observed in the ligated jugular veins, the authors failed to identify changes in the BBB permeability, neuroinflammation, demyelination, or clinical signs in the JVL group compared to the sham group. Thus, this murine model does not support CCSVI as a cause of demyelinating diseases such as MS.


As a matter of fact, MS lesions form around small venules. Histologically, the first step of the lesion formation is a lymphocyte cuffing initiating a perivenular inflammatory process. According to Zamboni’s theory, the signal that moves lymphocytes from blood to brain is iron which settles in MS lesions [35, 36]. Iron deposition in the brain, however, is observed in other “neurodegenerative” diseases, such as Parkinson’s disease [37] and Alzheimer’s disease (AD) [38], without evidence of MS-like inflammatory plaque formation. Moreover, the childhood disorders characterized by “neurodegeneration with brain iron accumulation” [39] have neither the clinical nor the magnetic resonance features of MS.

Recently, in an animal model of brain inflammation (called cerebral experimental autoimmune encephalomyelitis) that mimicked MS, iron deposition around vessels was found to occur independently of inflammation, providing evidence against the hypothesis that iron deposits necessarily account for inflammatory cell infiltrates in MS [40]. However, it has to be pointed out that histological data from pathological disorders occurring in other tissues (i.e., iron deposition in the liver [41] or in synovial fluid of arthritic joints [42] suggest that iron might play different roles in the loco-regional mechanisms of inflammation and/or degeneration, and metal ions have been implicated in the etiology of neurodegenerative disorders [43]. As mentioned above, dysregulation of brain iron and copper homeostasis was suggested to be a key factor in early neuropathological events in AD, including oxidative stress, amyloid deposition, tau phosphorylation, and neuronal cell cycle regulatory failure, leading to apoptosis [44]. Thus, we cannot exclude that iron deposition following inflammation may contribute to the neurodegenerative process that characterizes the progressive forms of MS. Indeed, variable degrees of iron deposition within MS lesions was described almost one century ago, although the histological data appear quite dyshomogeneous [45, 46]. Adams et al. found perivenular and perilesional hemosiderin deposits in a subset of 30 % of 70 MS-patients and attributed their finding to suspected microhemorrhage due to changes in permeability of vascular walls caused by local inflammation [27]. Although in a murine model of experimental autoimmune encephalomyelitis the extravasation of erythrocytes and iron deposits in cells of the macrophage/microglia lineage were observed, erythrocyte extravasation is far to be a common feature in MS [47]. Finally, recent studies have found elevated iron stores in the deep grey matter of MS patients not related to inflammatory plaques [4850]. Considering that basal ganglia and thalamus are hypotrophic in some MS patients and this has been correlated to some MS-related symptoms, such as fatigue [5153], it might be possible that iron deposition in some “critical” brain areas may contribute to the complex and still inexplicable clinical picture of MS.

From a methodological point of view, it has to be pointed out that several magnetic resonance imaging methods have been used to demonstrate and measure iron deposition in the brain, including conventional relaxation rate measures (R2*) and unconventional field-dependent relaxivity increase, phased and susceptibility-weighted imaging, direct saturation imaging, and quantitative susceptibility mapping [54]. These various imaging approaches are in effect quantifying different aspects of brain iron deposition, making quite difficult to compare the literature data and to draw conclusions. Moreover: (1) iron assessment in the white matter is particularly challenging because of its very low physiological concentration in this compartment (under the detection level of most methods), (2) iron is distributed in a heterogeneous fashion among different brain regions and cells, and these concentrations are not static, (3) abnormal iron accumulation in the diseased brain areas and, in some cases, alterations in iron-related proteins have been reported in many neurodegenerative diseases, (4) these diseases are characterized by a great clinical, neuroimaging, and histological heterogeneity [55, 56]. Thus, it is quite surprising to see the simplicity with which “iron accumulation in the brain” is used to support pathological theories for MS.

Finally, the cause-effect relationship between iron deposition and inflammation was also challenged by the analysis of CSF ferritin levels, believed to correlate with the degree of CNS iron load. RRMS and PPMS patients had normal ferritin concentrations, while SPMS had significantly increased CSF ferritin [57]. These findings indicate that iron deposition may occur in a subset of MS patients but is probably secondary to inflammation. Indeed, according to the CCSVI hypothesis, we would expect high ferritin levels in early disease stages, namely, when inflammation starts or inflammatory lesions show activity. Thus, the available data do not support a role for iron in triggering lymphocyte crossing through the BBB.


Fibrin plays a central role in the CCSVI theory: Zamboni et al. usually cite data from the literature indicating that fibrin deposition represents one of the early pathological events in MS brain. They state that demonstration of fibrin deposition has been identified as the early evidence of microvascular injury in “acute” MS. Indeed, Wakefield et al. studied brain specimens obtained from three cases of “acute” MS and found that the mural architecture of veins and capillaries was normal, while fibrin deposition was evident not only in inflammatory areas but also in areas in which myelin was preserved and reactive changes in the cerebral parenchyma were absent. Fibrin deposition was not seen in brain specimens of controls [29]. However, the three cases described by these authors were quite peculiar: one case was a fatal acute disseminated encephalomyelitis (ADEM) in a 14-year-old boy and two cases had huge tumor-like inflammatory lesions that underwent craniotomy. Clearly, these three cases cannot be considered “typical” cases of MS and, thus, the suggestion that “…focal endothelial cell activation which progresses to occlusive vascular inflammation is a precursor of both cellular infiltration of vessels and demyelination” should be taken with extreme caution.

In his attempt to demonstrate similarities between the histological aspects of chronic venous disorders (CVDs) and MS, Zamboni often cites the histopathological studies performed by Adams. Curiously, these citations are clearly inappropriate. For instance, in the paper published in Phlebology 2010 [20], Zamboni and Galeotti stated “Interestingly, pericapillary fibrin cuffs, a well-known marker of venous hypertension, have also been demonstrated in CCSVI associated with MS [2, 26, 27, 57].” Reference [2] is Zamboni’s manuscript on J R Soc Med 2006 (i.e., the big idea), while Refs. [26, 27, 57] are three papers by CWM Adams, who never studied CCSVI. Indeed, in the remarkable work “Pathology, histochemistry and immunocytochemistry of lesions in acute multiple sclerosis” (J Neurol Sci 1989, cited by Zamboni in several papers), Adams et al. studied 20 patients with “acute MS” who died within 6 months after clinical onset or after a dramatic relapse. However, if one reads this paper accurately and analyzes the clinical data, four patients had Marburg’s disease, one had Balo’s concentric sclerosis, while four probably had neuromyelitis optica (lesions only in the optic nerve and/or in the spinal cord). Staining with anti-fibrinogen was positive in 8, negative in 5 and not available in 7, and was more pronounced in cases with a fulminant course (<3 months), suggesting “…that fibrinogen leakage is an index of increased vascular permeability in acute lesions” [58]. In the same above-mentioned paper, Zamboni and Galeotti state “Raised venous pressure can stretch vein walls sufficiently to separate the tight junctions between endothelial cells forming the blood–brain barrier” and, to strengthen this statement, they cite a work on pulmonary capillaries [59]. Moreover, they cite the paper by Singh and Zamboni to affirm that “…an impressive parallel has been delineated between the inflammatory process activated in course of CVD, and that profoundly studied in MS” [3]. Unfortunately, that paper too is a review of literature data, often cited out of place, with no scientific and reproducible observations.

Worth of interest, however, are recent data in human diseases and in animal models of neurodegeneration that have extended the possible role of fibrin in nervous system pathology, suggesting that this protein (that is not produced by CNS cells) is a component of the perivascular extracellular matrix that regulates inflammatory and regenerative cellular responses in neurodegenerative diseases [60]. Moreover, a fibrin-derived peptide (γ377–395), that blocks fibrinogen-Mac-1 interactions, was found to attenuate microglia activation and suppress relapsing paralysis in an animal model of CNS inflammation [61], thus suggesting that targeting the γ377–395 epitope could represent a potential therapeutic strategy for MS. Also this, however, has nothing to do with the CCSVI theory.

In brief, no scientific data supports Zamboni’s speculation on fibrin. The possible role played by fibrin within the inflammatory network is probably more complicated than previously supposed and further studies are needed to clarify the role played by this protein in neurological diseases.

Brain perfusion

The demonstration of reduced brain perfusion in MS patients has also been suggested to speak in favour of the CCSVI hypothesis. Unfortunately, literature data appear quite contradictory and their interpretation is not univocal. Brain perfusion can be studied by applying dynamic susceptibility contrast-MRI (DSC-MRI). In brief, this technique is based on the acquisition of a set of fast T2*-weighted images before, during and after the injection of gadolinium-DTPA, which reduces the transverse tissue relaxation time (T2*) and, thus, the MRI signal. DSC-MRI allows the analysis of several perfusion parameters, namely the cerebral blood flow (CBF), the cerebral blood volume (CBV) and the mean transit time (MTT) [6264]. When applied to MS, DSC-MRI gave conflicting results. Compared to normal controls (NC), patients with relapsing-remitting MS (RRMS) were found to have a reduced CBF, normal or reduced CBV, and increased or unchanged MTT in the normal appearing white matter (NAWM) [65, 66]. When WM lesions were classified on the basis of their post-contrast T1 image intensities, acute plaques had significantly higher blood volumes than NAWM, while chronic plaques had lower CBV, indicating that active MS lesions were accompanied by vasodilation while inactive lesions by a decreased perfusion [67].

Early changes in the regional perfusion preceding evidence of BBB breakdown and inflammation were also described in NAWM [68, 69]. Moreover, in apparently inactive lesions the CBV was found either increased or reduced, and the MTT was prolonged compared to the NAWM of NC [68]. When studying the perfusion of cortical lesions, we also observed a heterogeneous behavior, with some lesions having increased perfusion parameters. The longitudinal follow-up of these lesions showed a progressive decrease of the perfusion parameters, further indicating that the perfusion is increased in “active” inflammatory lesions and decreased in chronic lesions (personal data: manuscript in preparation). Taken all together, the perfusion data speak against the hypothesis that a venous reflux is the primum movens of MS pathology. With disease progression, the accumulation of sclerotic lesions and the reduced frequency of new inflammatory lesion would determine a global reduction of MS brain perfusion that probably involves the areas of normal appearing white matter adjacent to lesions. Thus, the CCSVI theory is not supported by the perfusion studies.

The ultrasound puzzle: can order spring from chaos?

Several studies reporting data that strongly disagree with the CCSVI theory are currently available in PubMed (see Table 1). None of these studies, performed with different methodologies, was able to find a clear relationship between abnormal venous outflow and MS.
Table 1

Summary of the percentage of CCSVI-positive MS patients according to the clinical subtypes









Auriel et al. [70]



Bastianello et al. [72]

82 %

91 %

92 %

Baracchini et al. [6]




Baracchini et al. [7]




Centonze et al. [8]

36/69 (52 %)

6/15 (40 %)

20/56 (36 %)

Doepp et al. [9]




Doepp et al. [10]



Krogias et al. [11]




Mayer et al. [12, 13]




Sundström et al. [15]



Tanaka et al. [16]



Tsivgoulis et al. [71]




Wattjes et al. [17]




Zivadinov et al. [19]

8/21 (38.2 %)

94/191 (49.2 %)

58/80 (62.5 %)

6/11 (54.5 %)

11/26 (42.3 %)

37/163 (22.7 %)

Yamout et al. [18]

1/11 (9 %)


Zamboni et al. [1, 4]

33/35 (94 %)

20/20 (100 %)

10/10 (100 %)

0/93 (0 %)

0/142 (0 %)

CIS clinical-isolated syndrome, SP secondary progressive, PP primary progressive, OND other neurological disorders, NC normal control

Using Doppler criteria three German studies found no evidence of CCSVI, respectively in 56 (41 RR and 15 SP) [9], 20 (17 RR and 3 SP) [12] and 10 (two RR, seven SP and one PP) [11] MS patients and in a total number of 40 NC. In two more recent studies (from Israel and Japan, respectively), one in 27 MS patients and 32 NC and one in 17 RRMS and 11 patients with neuromyelitis optica, completely negative results have also been reported [16, 70]. In a Greek study, a series of 42 MS and 43 NC was analyzed and a reflux in internal jugular vein (IJV) was documented in one patient and one control subject, both in sitting and supine posture during apnea, and following Valsalva maneuver (VM), the presence of IJV valve incompetence was documented in three MS (7 %) and four NC (9 %; p > 0.999), but no patient or control had evidence of CCSVI [71]. Thus, a normal venous outflow was found in different ethnic populations.

A recent Italo-Canadian study involving six centers, with serious methodological limitations (lack of healthy controls, not blinded) has clearly disclosed an unacceptably high between-center variability, especially in single CCSVI criteria (i.e., the percentage of positivity for criterium 2 ranged from 29 to 91, for criterium 5 ranged from 5 to 86 %), thus making the conclusions of the study difficult to accept [72].

Using magnetic resonance venography (MRV) and flow quantification study, no evidence of venous backflow was observed in 20 patients with MS (19 RR, 1 pp) and 20 NC [15], and in a case–control study, no differences in IJV outflow, aqueductal cerebrospinal fluid flow, or the presence of IJV blood reflux could be demonstrated in patients with MS [17]. Moreover, no significant difference in morphological features of flow in the IJV and VV were found between 57 patients (41 RRMS and 16 SPMS) and 21 NC by means of 3-T MRV [10]. Recently, 40 MS patients underwent contrast-enhanced MRV for assessment of IJV and azygos vein (AV) narrowing (graded into three groups: 0–50 %, 51–80 %, and >80 %) and ECCS analysis of blood flow direction, cross-sectional area, and blood volume flow in both IJV and vertebral veins (VV) [19]. The findings of this study strongly contradict the postulated 100 % prevalence of CCSVI criteria in MS, and suggest that MRV seems more sensitive to detect IJV narrowing compared to ECCS. However, a measurable hemodynamic effect was observed only in vessel narrowing >80 %. Others, however, had quite opposite results: Doppler sonography was found to be more sensitive than MRV in detecting intraluminal structural and functional venous abnormalities in patients with MS compared with healthy controls [73]. In this study, however, structural abnormalities of the IJVs were found in 54 % NC and 74 % MS, while functional abnormalities were observed in 33.3 % NC and 54.7 % MS.

When a selective extracranial venography was performed on 42 patients with MS, extracranial venous stenosis was seen in 7/29 patients with early MS and 12/13 with late MS. Only three of 42 patients (all in the late MS group) demonstrated two vessel stenoses [18], suggesting a relationship between CCSVI and secondary progressive MS.

Strictly applying Zamboni’s ultrasound criteria and methodology, variable percentages of CCSVI-positive MS and NC were found by others groups. Zivadinov et al. [19] found evidence of CCSVI in 49.2 % RR, 54.5 % PP and 62.5 % MS patients, but also in 22.7 % NC. In patients with clinically isolated syndromes (CIS) suggestive of MS, CCSVI was demonstrated in 38 % of the cases. This finding and the observation of the increased percentage of CCSVI with disease duration conducted these authors to the speculation that CCSVI could be caused by MS and not vice versa [19]. On the contrary, Centonze et al. [8] found a higher percentage of CCSVI-positive patients in the RRMS group (52 %) than in the SPMS group (40 %), but no difference with the NC (36 %) was observed. These results indicated that CCSVI had a role neither in MS risk nor in MS severity. However, the demonstration of CCSVI in a relevant percentage (1/3–1/4) of normal individuals raises several questions: (1) If CCSVI is not the cause of MS and can be demonstrated in NC, what is its pathological significance? (2) If in MS-affected individuals MS is the cause of CCSVI, what determines the appearance of CCSVI in normal individuals? (3) How to manage a potentially dangerous syndrome that affects more than 1 billion human beings? (4) Might it be that CCSVI is a fascinating chimera that does not exist in the real world?

To answer these questions, we performed two large studies aimed at analyzing the occurence of CCSVI at MS onset, to elucidate the possible causative role of CCSVI in MS and in the progressive forms of MS, to investigate as to whether CCSVI could play a role in determining disease progression [6, 7]. Fifty patients presenting with a clinically isolated syndromes and having evidence of dissemination in space of the inflammatory lesions underwent a comprehensive diagnostic workup, including extracranial (ECDS) and transcranial venous echo-color Doppler sonography (TCDS). Patients who showed evidence of CCSVI were further evaluated by selective venography. Fifty MS-matched NC, 60 patients with transient global amnesia (TGA), and 60 TGA-matched NC were studied. TCDS was normal in all patients with CIS. One or more abnormal extracranial venous echo-color Doppler findings were observed in 26 of 50 (52.0 %) of the patients with CIS, 35 of 110 (31.8 %) of the controls and 41 of 60 (68.3 %) of the patients with TGA. The eight (16 %) patients with CIS who fulfilled the diagnosis of CCSVI were further evaluated blindly by selective venography, which did not disclose any venous anomalies. Thus, we could not demonstrate any causative effect of CCSVI on MS [6]. In the second study, we analyzed 60 patients with chronic progressive form of MS (35 SP, 25 PP) [7] and 60 age-/gender-matched normal controls (NC). TCDS was normal in all patients. ECDS showed one or more abnormal findings in 9/60 (15.0 %) patients (7/35 (20.0 %) SPMS, 2/25 (8.0 %) PPMS) and in 14/60 (23.3 %) NC (p not significant for all comparisons). CCSVI criteria were fulfilled in 0 NC and 4 (6.7 %) MS patients: three SPMS and one PPMS. VGF, performed, blindly, in 6/9 patients, was abnormal only in one case who had bilateral internal jugular vein stenosis. These findings indicate that CCSVI is not a late secondary phenomenon of MS and is not responsible for disability progression.

CCSVI and ultrasound criteria

The five ultrasonographic criteria suggested by Zamboni constitute the backbone of CCSVI diagnosis. Surprisingly, a very close look at these criteria discloses a dangerous mix of miscitation, manipulation of known data and methodological errors.

For criterion 1 (reflux in extracranial veins) Zamboni et al. [1] used the threshold value of 0.88 s to discriminate IJV and VV physiological back flow due to valve closure from pathological reflux without performing the VM and they found that 71 % of MS patients had a pathological reflux versus 0 % of controls. However, this threshold value comes from a totally different study on IJV valve insufficiency during a controlled VM [74] where it was chosen to differentiate VM-induced insufficiency through insufficient valves lasting >1.23 s, from physiological backward flows during normal valve closure, lasting 0.22–0.78 s. In this study it was found that about 30 % of normal subjects have a physiological (t < 0.88 s) back flow during normal valve closure. Furthermore, the utilization of this threshold by Zamboni for assessing reflux in vertebral veins, other than IJV valve insufficiency, is also scientifically incorrect. Finally, the presence of a reflux >0.88 s in the internal jugular vein is more likely to indicate IJVI rather than MS (Table 2).
Table 2

Ultrasound CCSVI criteria

1. Reflux (t > 0.88 s) in the IJVs and/or in the VVs in sitting and supine position

2. Reflux (t > 0.5 s) in the DCVs

3. High-resolution B-mode evidence of proximal IJV stenoses (CSA ≤ 0.3 cm2)

4. Flow not Doppler-detectable in the IJVs and/or VVs despite numerous deep inspirations

5. Reverted postural control of the main cerebral venous outflow pathways: a missing increase of IJV diameter in the supine position

CSA cross-sectional area of the internal jugular vein, IJV internal jugular vein, VV vertebral vein, DCVs deep cerebral veins

For criterion 2 (reflux in the deep cerebral veins), the intracranial veins and sinuses were not examined through the transtemporal bone window for which there are published ultrasound criteria and velocity data [75, 76]. Zamboni et al. used a new bone window (supracondylar) for which there are neither accepted published criteria nor normative data, and the figures published are not compatible with normal anatomy [1]. With regards to cerebral venous reflux, they found this in 61 % of MS patients; however, this evaluation requires a Doppler spectrum analysis, because a color-based approach is inadequate and can easily lead to the misinterpretation of flow direction. More importantly, the rationale of adopting a threshold value of 0.5 s to discriminate pathological reflux in the deep cerebral veins is unclear. This value was derived from studies in the veins of the leg where it served to quantify venous valve insufficiency following deflation of a tourniquet [77, 78]. The rationale for transferring this value from the legs to the brain is very questionable since it has never been validated for deep cerebral veins. The validity and significance of data collected by this method is therefore unclear especially if it is used to diagnose CCSVI, where cerebral reflux is not described by the same author as associated with valve incompetence.

Criterion 3 (proximal IJV stenosis) is also based on a scientifically incorrect application of data obtained in a different setting: it defines a proximal stenosis of the IJV as a cross-sectional area (CSA) in the recumbent position ≤0.3 cm2 [1]. This cut-off value was derived from a study on intensive care patients [79], with possible confounders such as mechanical ventilation and hypovolemia. It can therefore not be used as a reference point in healthy subjects. Furthermore, the original authors [79] still reported a CAS ≤0.3 cm2 in 20 % of their patients. It is difficult to decide where to measure the diameter of the vein since IJVs are normally tortuous and the most proximal and distal parts near the superior and inferior bulb are physiologically dilated more than others. It is important to stress that even mild pressure exerted by the ultrasound probe or by a contraction of the cervical musculature itself can alter the diameter of the vein leading to false-positive results.

Criterion 4 (absence of color-coded flow signals) has never been validated before: the inability to detect flow in the IJVs and/or in the VVs during deep inspiration, according to Zamboni et al. provides indirect evidence of venous obstruction [1]. In the original paper flow was assessed at rest, rather than during deep inspiration, and this finding was never discussed in the context of venous obstruction [80]. Moreover, a lack of flow is not necessarily due to obstruction since it can occur e.g., at 15° in both IJVs in healthy subjects [76]. In the upright position, there is a dramatic reduction and frequently a complete cessation of blood flow in the IJV. In the supine position there may also be no flow in the VVs [77]. Furthermore, an inadequate setting of ultrasound indices such as pulse repetition frequency might lead to an apparent absence of color-coded signal and a misinterpretation of no flow.

The fifth criterion (reverted postural flow) examines the presence of a physiological shift of cerebral venous drainage from the jugular venous system to the vertebral plexus with postural change: from the supine to the sitting position. In normal subjects, subtracting the CSA measured in the supine position from that in a sitting position (ΔCSA) is usually negative [76]. Instead, Zamboni wrongly considered that a negative ΔCSA value would represent a reverted postural control of the main cerebral venous outflow pathways [1]. Furthermore, similar to criterion three, a mild pressure exerted by the ultrasound probe or by a contraction of the cervical muscles may alter the diameter of the vein possibly leading to false-positive results. A more correct method would be to calculate the difference of blood flow (CSA × velocity) in the two positions (supine and sitting) as has been recently performed not confirming the hypothesis of Zamboni et al. [10, 15, 17, 19].

A very important issue is the cut-off point of these criteria to diagnose CCSVI. In fact, it is unclear how Zamboni decided that two or more of the five ultrasound criteria may be used to diagnose CCSVI. Diagnostic criteria using a new alternative method (i.e., ultrasound) are usually compared with a validated gold-standard investigation (venography according to Zamboni et al.). However, Zamboni et al’s comparison of venography in 65 CCSVI ultrasound-positive MS patients was not blinded and is therefore open to bias. There was also no validation of the CCSVI criteria by different and independent observers. Finally, subsequent studies using MR-venography could not confirm differences regarding cerebrospinal drainage in MS patients and controls [10, 15, 17, 19].

Endovascular treatment data

No study on the benefit of angioplasty that meets the criteria of evidence-based medicine (multicenter, randomized, controlled) has been performed in MS patients with CCSVI. The first interventional study [4] did not meet minimum requirements of scientific accuracy (blinding, sham-operation and control group design). Moreover, the recent interventional study by the same group who claimed “to address these criticisms” [80] had the same weaknesses of the former trial. A recent open-label study on 60 patients who were “liberated” and then followed for 6 months gave substantially negative results [81]. We are sure that Zamboni et al. agree with the scientific community that studies ignoring the rules of evidence-based medicine and good clinical practice introduce such relevant biases and an unavoidable placebo effect that invalidate any evidence of a possible effectiveness of endovascular treatment in MS. In other words, not only the scientific community (this should not be questioned) but also MS patients and their relatives as well as media operators should be aware that the CCSVI theory lacks meaning and fact, and is based on the negation of the basic principles of science, since the evidence of endovascular treatment efficacy cannot be based on the awareness of both the evaluating physician and the evaluated subject and on the complete lack of blinding [13, 82].

The generally well-tolerated technique of endovascular intervention naturally bears a certain risk of complication [8385], which further increases when stent implantation requires platelet-inhibiting drugs. In a recent report on 240 MS patients who underwent 257 endovascular treatments, post-angioplasty or post-procedural (<30 days) venous thrombosis was observed in 11.3 % of the cases, sustained cardiac arrhythmias that required admission in 1.2 %, while one patient developed a stress-induced cardiomyopathy [86]. In a series of 331 MS patients with CCSVI undergoing 344 interventions (angioplasty in 192 cases and stent placement in at least one vein in 152 cases) [87, 88], complications included early stent thrombosis (1.2 %), difficulty with removal of the angioplasty balloon or delivery system (1.5 %) with one patient requiring a surgical cutdown to remove an angioplasty balloon, transient atrial fibrillation (0.6 %), and local bleeding from the groin (1.2 %) with two patients having a pseudo-aneurysm requiring a thrombin injection for treatment. These two studies were not designed to assess the clinical outcomes after endovascular treatment, and, curiously, no neurologist is listed among the authors of the papers.

The Society of Interventional Radiology (SIR) and the Canadian Interventional Radiology Association agreed on a position statement regarding endovascular intervention specific to CCSVI [89]: the available literature was considered inconclusive regarding the pathogenetic role of CCSVI in MS and the efficacy of endovascular intervention. The Cardiovascular and Interventional Radiological Society of Europe (CIRSE) clearly stated that, due to limited and controversial studies, the endovascular procedures could not be considered to bring more benefit than harm and were thus not recommended to any MS patient outside of a proper clinical trial [90]. This view is also shared by patients’ organizations, such as the US National MS Society and the Italian Society for Multiple Sclerosis.

Moreover, the practice statements published by the European Federation of the Neurological Societies and European Neurological Society Multiple Sclerosis Scientist Panel and European Committee for Therapy and Research In Multiple Sclerosis Executive Committee [91], should be carefully considered given the prominence accorded to lay discussions regarding the CCSVI hypothesis in the non-scientific journals, internet blogs and social networks, leading to the expansion of an unethical lucrative industry of equivocal “CCSVI-specialized centres” who are eager to respond to patient demands by offering them diagnosis and/or endovascular treatments at costs ranging from 5,000 to 10,000 € (in Italy), without any controlled scientific evidence.


The inconsistency of the CCSVI theory is self-explanatory. No piece of the CCSVI puzzle has a solid supportive scientific evidence; CCSVI appears to be a rather alien condition and its existence should be definitely questioned. Moreover, no proven (i.e., based on strict scientific methodology) therapeutic effect of the “liberation” procedure has been shown up to date. In this chaos of data lacking true scientific evidence of benefit, even a low peri-procedural risk has to be considered unethical by all responsible clinicians.

Therefore, in our opinion routine endovascular treatment of CCSVI in MS has to be firmly inhibited, while controlled multicentre “liberation” trials should not be allowed until a solid scientific evidence of a causal relationship between CCSVI and MS has been clearly demonstrated.

Conflict of interest

Dr. C. Baracchini serves on the executive committee of the European Societyof Neurosonology and Cerebral Hemodynamics; has received funding for travel and speaker honoraria from Pfizer-Inc, Sanofi-Aventis, Laboratori-Guidotti S.P.A. and Novartis; serves as an Associate Editor for BMC Neurology. Dr. M. Atzori reports no conflict of interests. Prof. P. Gallo serves on scientific advisory boards for and has received funding for travel or speaker honoraria from Biogen-Idec, Merck-Serono, Bayer-Schering-Pharma, Sanofi-Aventis, and Novartis; and receives research support fromBiogen-Idec, Merck-Serono, Bayer-Schering-Pharma, Sanofi-Aventis, Novartis and the Italian Ministry of Public Health.

Copyright information

© Springer-Verlag 2012