“Liberation treatment” for chronic cerebrospinal venous insufficiency in multiple sclerosis: the truth will set you free

Background Chronic cerebrospinal venous insufficiency (CCSVI) has recently been introduced as a chronic state of impaired cerebral or cervical venous drainage that may be causally implicated in multiple sclerosis (MS) pathogenesis. Moreover, percutaneous transluminal angioplasty of extracranial veins termed “Liberation treatment” has been proposed (based on nonrandomized data) as an alternative therapy for MS. Methods A comprehensive literature search was conducted to identify available published, peer-reviewed, clinical studies evaluating (1) the association of CCSVI with MS, (2) the reproducibility of proposed ultrasound criteria for CCSVI detection (3) the safety and efficacy of “Liberation treatment” in open-label and randomized-controlled trial (RCT) settings. Results There is substantial heterogeneity between ultrasound case–control studies investigating the association of CCSVI and MS. The majority of independent investigators failed to reproduce the initially reported high prevalence rates of CCSVI in MS. The prevalence of extracranial venous stenoses evaluated by other neuroimaging modalities (contrast or MR venography) is similarly low in MS patients and healthy individuals. One small RCT failed to document any benefit in MS patients with CCSVI receiving “Liberation treatment”, while an exacerbation of disease activity was observed. “Liberation treatment” has been complicated by serious adverse events (SAEs) in open-label studies (e.g., stroke, internal jugular vein thrombosis, stent migration, hydrocephalus). Conclusion CCSVI appears to be a poorly reproducible and clinically irrelevant sonographic construct. “Liberation treatment” has no proven efficacy, may exacerbate underlying disease activity and has been complicated with SAEs. “Liberation treatment” should stop being offered to MS patients even in the settings of RCTs.


Introduction
Multiple sclerosis (MS) is a chronic, primary inflammatory disease of the central nervous system (Gold et al. 2006). In 2009, a new concept for the pathogenesis of MS based on the idea of an impaired cerebrospinal venous drainage was postulated, differing from the established concept of the multifactorial pathogenesis of MS (Zamboni et al. 2009a). This hypothesis was based on specifically developed ultrasound features which were stated to detect cervical or cerebral abnormalities of venous drainage leading to increased intracranial venous pressure, subsequently followed by blood-brain barrier breakdown causing iron deposition in brain parenchyma initiating the development of MS (Zamboni 2006). Moreover, percutaneous transluminal angioplasty of extracranial veins (termed "Liberation treatment") has been proposed (based on nonrandomized data) as an alternative therapy for MS by the same group of investigators introducing CCSVI hypothesis (Zamboni et al. 2009c). Despite the lack of higher class evidence "Liberation treatment" has gained a considerable amount of attention and emotional involvement by MS patients worldwide (Chafe et al. 2011) and has started to be offered as a potential therapeutic option in MS patients in nonrandomized and uncontrolled studies (Hubbard et al. 2012;Mandato et al. 2012;Ghezzi et al. 2013b).
However, numerous independent investigators failed to detect any association between CCSVI neurosonology criteria and MS in numerous case-control studies, while "Liberation treatment" has been complicated with serious adverse events (SAEs) leading to substantial criticism of "venous hypothesis" of MS pathogenesis (Barkhof and Wattjes 2013;Valdueza et al. 2013).

Methods
In view of the former considerations, we conducted a comprehensive literature search to identify available published, peer-reviewed, clinical studies evaluating (1) the association of CCSVI with MS using different ultrasound modalities, (2) the reproducibility of proposed ultrasound criteria for CCSVI detection, (3) the safety and efficacy of "Liberation treatment" in open-label and randomizedcontrolled trial (RCT) settings.
Our literature search through MEDLINE was based on the combination of terms "Chronic cerebro-spinal venous insufficiency", "multiple sclerosis", "transcranial sonography", "iron", "ultrasound", "Liberation treatment," and "venous angioplasty". Last literature search was conducted on 14 August, 2014. Reference lists of all articles that met the criteria and of relevant review articles were examined to identify studies that may have been missed by the database search. Titles, abstracts and, whenever appropriate, full texts of all identified studies were screened independently by two reviewers in English (GT, CK) and two reviewers in German (SM, CK) journals. Potential disagreements were resolved by consensus of all contributing authors. Duplicate publications and publications in other than the English or German language were excluded from further evaluation.

Results
The venous hypothesis of the pathogenesis of MS The "venous hypothesis" postulates that disturbances of the venous drain from the cervical and/or spinal venous system leads to a congestion and increase in the intracranial venous pressure (Singh and Zamboni 2009). Given his experience in vascular surgery, Dr Zamboni who introduced the "CCSVI hypothesis" saw pathophysiological parallels between a chronic state of cervical or cerebral venous insufficiency and MS (Zamboni 2006). Similar to the chronic venous insufficiency (CVI) in the leg veins, he postulated that cerebral venous stasis may induce endothelial damage leading to blood-brain barrier disruption, which in turn may result in extravasation of erythrocytes (Zamboni 2006). He also hypothesized that disintegrating erythrocytes may be dismounted by macrophages which may cause a local deposition of iron (Ackermann et al. 1988). The focal increased amount of iron may in turn induce a chronic inflammatory reaction with an upregulation of the migration of leukocytes in the subcutaneous matrix via expression of adhesion molecules (ICAM, VCAM) and of selectins (Colleridge-Smith et al. 1988). He underlined that macrophages and T-lymphocytes will play a predominant role in this process, since macrophages have been shown to phagocyte the accumulated iron in subcutaneous tissue and store it intracellularly (Wilkinson et al. 1993;Takase et al. 2004). Moreover, the CCSVI hypothesis underscores that extracellular depositions and iron-loaded macrophages are not only the histopathological feature of CVI but can also be found in MS plaques (Adams et al. 1989) and have been shown to stimulate the immune system (Weilbach et al. 2004). The fact that MS-plaques are located in the perivenous region led to the hypothesis that the pathophysiology of MS may be mediated through a chronic inflammatory reaction whose cause lays in the impaired venous outflow (as well as in CVI).
This hypothesis acquired a high resonance in some patient groups as well as in the media, as it was presented as a comprehensible, potentially curable cause of MS which was easy to understand (Pullman et al. 2013). Moreover, patient organisations and advocacy groups, scientific societies, and health-care authorities started funding CCSVI at an International level (Canada, Italy, United States) under pressure from the media and the blogosphere (Pullman et al. 2013).

External validation of ultrasound criteria for CCSVI detection
Zamboni et al. reported in their pivotal study that neurosonology had 100% accuracy parameters (sensitivity, specificity, positive predictive value, negative predictive value) to discriminate MS patients from Healthy Controls (HC) using a set of ultrasound criteria developed to detect impaired cervical or venous drainage (Zamboni et al. 2009b). Consequently, they introduced evidence of two positive out of five proposed ultrasound criteria as necessary condition for CCSVI diagnosis (Table 1, Fig. 1) (Zamboni et al. 2009a). Moreover, they recommended a detailed neurosonology protocol for CCSVI screening (Nicolaides et al. 2011). However, it should be noted that blinding of sonographers was suboptimal in the majority of studies of Zamboni's group.
A series of studies conducted by independent investigators failed to reproduce the findings reported by Zamboni and coworkers in their pivotal studies. In a first ad-hoc investigation of a small unselected group of MS patients in Bochum, the prevalence of CCSVI was similar in MS patients (20%) and HC (10%) (Krogias et al. 2010). In a larger investigation conducted in Berlin, CCSVI criteria were not met in a single individual out of 56 investigated patients (Doepp et al. 2010). In a North-American study using sonographers trained by Zamboni, a higher prevalence of CCSVI was documented in MS patients (63%) than in HC (26%) (Zivadinov et al. 2011). Moreover, a post hoc analysis indicated that CCSVI was independently associated with a more progressive MS course. The association of CCSVI and MS was also reproduced by a Jordanian (Al-Omari and Rousan 2010) and a Polish (Zaniewski et al. 2013) group of investigators that were also offering "Liberation treatment" in their MS patients.
In contrast to the previous observations, an Italian study, investigating only MS patients with clinically isolated syndrome, reported a normal ultrasound investigation in 84% of study population (Baracchini et al. 2011). Moreover, CCSVI could not be detected in additional series of MS patients investigated in Greece (Tsivgoulis et al. 2011) and Germany (Frankfurt/Giessen) (Mayer et al. 2011). In addition, the largest to date, methodologically robust (using both local and central blinded readers) ultrasound case-control study involving 1874 subjects from 35 Italian centers reported a similar (very low) prevalence of CCSVI in MS (3%) and HC (2%) (Comi et al. 2013). Interestingly, the poor interrater and intrarater agreement in CCSVI ultrasound criteria reported both by Italian (Comi et al. 2013) and Greek (Tsivgoulis et al. 2011) investigators underscore the lack of reproducibility of the proposed neurosonology protocol (Table 2). More specifically, the positive agreement between central and local readers in the Italian study was disappointedly low (18%) (Comi et al. 2013). Finally, in an assessor-blinded, case-control, multicentre Canadian study using doppler ultrasound equipment identical to that used in the pivotal study by Zamboni et al. (2009a) and experienced sonographers trained in the center of Zamboni in Ferrara, the prevalence of CCSVI was similar in MS (44%) and HC (45%) (Traboulsee et al. 2014).
In view of the discrepant results between different investigators, recent meta-analyses Krogias et al. 2013;Zwischenberger et al. 2013;Tsivgoulis et al. 2014) have suggested an independent association between Table 1. Proposed ultrasound criteria for CCSVI diagnosis (at least two criteria present).   an ultrasound-based diagnosis of CCSVI and MS with OR ranging between 1.9 and 13.5. However, considerable heterogeneity (I squared statistic >50%) across included studies was documented Krogias et al. 2013;Tsivgoulis et al. 2014), while a factor contributing to this heterogeneity (according to sensitivity analyses of the largest to date metaanalysis) appears to be the involvement of investigators in endovascular procedures supporting "Liberation treatment" as a novel therapeutic strategy for MS (Tsivgoulis et al. 2014).
Evaluation of cerebral venous drainage in MS using other than ultrasound neuroimaging modalities In contrast to the CCSVI hypothesis, three Magnetic Resonance Imaging (MRI) studies failed to document a higher prevalence of impaired cerebral or cervical venous flow due to obstructions in cervical or thoracic veins in MS patients in comparison to controls (Sundstr€ om et al. 2010;Wattjes et al. 2011;Blinkenberg et al. 2012). Interestingly, a recent MR venography study failed to find increased prevalence of extracranial venous anomalies in children and adolescents with MS in comparison to healthy controls (Jurkiewicz et al. 2014). Moreover, a contrast venography study in 42 MS patients reported that extracranial venous stenosis is an unlikely cause of MS since it is not present in most patients early in the disease and rarely involves more than one extracranial vein (Yamout et al. 2010). It should also be noted that ultrasound diagnosis of CCSVI was not confirmed when the same patients were evaluated with another neuroimaging modality including Table 4. Methodological shortcomings of proposed neurosonology protocol for CCSVI diagnosis (Baracchini et al. 2011;Tsivgoulis et al. 2011;Valdueza et al. 2013).

Criterion
Methodological Brain and Behavior, doi: 10.1002/brb3.297 (7 of 12) magnetic resonance venography (Blinkenberg et al. 2012;Brod et al. 2013;Costello et al. 2014) or catheter venography (Baracchini et al. 2011;Traboulsee et al. 2014). More specifically, the sensitivity and specificity of the ultrasound criteria for detection of greater than 50% narrowing of any major vein on catheter venography was 40.6% (95% CI: 31.1-50.8%) and 64.3% (48.0-78.0%) in a recent multicenter, double-blinded (blinding of both sonographers and neuro-interventionalists) Canadian study. The highly discrepant results between ultrasound and other neuroimaging modalities lend support to the assumption that CCSVI may constitute a sonographic construct that is unlikely to contribute to MS pathogenesis (Filippi et al. 2011). Finally, Table 3 summarizes the findings of multimodal cross-sectional or case-control neuroimaging studies investigating the CCSVI hypothesis in MS pathogenesis. The majority of these studies failed to validate the CCSVI hypothesis using a combination of different sets of investigations including neurosonology, MR venography, and contrast venography.

Methodological shortcomings of proposed ultrasound protocol
The poor reproducibility of CCSVI diagnosis between and within sonographers as well as the low diagnostic yield of ultrasound against other neuroimaging modalities for detection of major cervical vein narrowing may be related to technical reasons including artificial compression of cervical veins by the ultrasound probe or contraction of cervical musculature leading to pseudostenosis, inappropriate selection of pulse repetition frequencies, misinterpretation of pulsation artifact from the adjacent carotid artery as venous reflux, failure to recognize intraluminal jugular septation causing IJV stenosis, misinterpretation of IJV valve insufficiency as IJV stenosis, inadequate patient compliance during sonographic evaluation of cervical veins at different body positions and during different phases of breathing (Baracchini et al. 2011;Tsivgoulis et al. 2011;Valdueza et al. 2013). Another plausible explanation may be associated with the incomplete blinding of the investigators and potential variabilities in the hydration status of MS patients (Comi et al. 2013). In addition, it is also clear that there are enormous variations in normal patterns of cerebral venous drainage within the healthy population, and that interpretation of patterns of venous drainage and venous obstruction can be highly subjective, subject to observer bias, and discrepant from institution to institution, depending upon the particular technology used for assessment as well as the expertise of the sonographers in cerebral and cervical venous ultrasound examinations (Valdueza et al. 2013).
The potential methodological shortcomings of the Table 5. Critique of published neurosonology images in pivotal studies introducing CCSVI hypothesis (Zamboni 2006;Zamboni et al. 2009a,b,c). Criterion proposed neurosonology protocol of CCSVI detection are summarized in Table 4. Finally, Table 5 displays a critical appraisal of available neurosonology images in the pivotal publications by Zamboni and colleagues introducing the CCSVI hypothesis.

Safety and efficacy of "Liberation treatment"
An argument postulated by several groups including patient advocacy groups, media representatives, and physicians is that even with suboptimal accuracy parameters a positive therapeutic effect of venous angioplasty and/or stenting cannot be excluded in MS patients with CCSVI constellations (Pullman et al. 2013;Zivadinov et al. 2013). Consequently, they postulate that access to such interventional therapies should not be refused in the context of randomized-controlled trials, while additional funding and intellectual energy are required to further investigate the venous hypothesis of MS pathogenesis (Zivadinov et al. 2013).
In contrast, there is growing literature underscoring that "Liberation treatment" for CCSVI can be complicated by serious adverse events including IJV or cerebral venous thrombosis, stent dislocation, vein dissection, femoral artery pseudoaneurysm, cranial nerve palsies, syncope or severe cardiac arrhythmias, hydrocephalus, and hemorrhagic complications of anticoagulation initiated following stent placement Ghezzi et al. 2013a;Tsivgoulis et al. 2014). Table 6 summarizes the reported complications of "Liberation treatment" across different studies.
Finally, a recently published class I sham-controlled, randomized, double-blind study investigating the safety and efficacy of venous angioplasty in MS patients fulfilling ultrasound criteria of CCSVI showed that venous angioplasty did not improve hemodynamic parameters in Table 6. Reported complications of "Liberation treatment" for treatment of CCSVI in multiple sclerosis patients.

Conclusions
The postulated hypothesis of a disturbed mechanical cervicospinal venous drain as a monofactorial etiopathogenic mechanism of MS should be discarded in view of the numerous independent external validation studies contradicting Zamboni's observations and the highly discrepant findings between ultrasound and other neuroimaging modalities. CCSVI appears to be a poorly reproducible and clinically irrelevant sonographic construct. "Liberation treatment" has no proven efficacy, may exacerbate underlying disease activity and has been complicated with serious adverse events. "Liberation treatment" should stop being offered to MS patients even in the settings of randomized-controlled trials, while further unnecessary expenditure of scarce funding resources needs to be discontinued. Physicians taking care of individuals with MS should spend time educating their patients with regard to the scientific evidence refuting CCSVI hypothesis as well as the potential complications and the lack of efficacy of "Liberation" treatment replacing the blogosphere as the main source of "sensational" but inaccurate information.