Does Chronic Cerebrospinal Venous Insufficiency Influence and Contribute to the Development of Multiple Sclerosis?
do not necessarily reflect the views of UKDiss.com.
Introduction
Multiple Sclerosis (MS) is a chronic inflammatory disorder of the central nervous system in which plaques of demyelination develop in the brain and spinal cord1. The disease can cause varied symptoms and signs of neurological dysfunction including optic neuritis, gait ataxia, intention tremor, vertigo and transient weakness2. Peak onset is at around thirty years old, and in the early stages of the disease symptoms can occur and disappear unpredictably, creating a relapsing-remitting (RRMS) disease pattern. Over time however, symptoms may become prolonged and more severe as the patient enters a primary progressive (PPMS) and then a secondary progressive (SPMS) phase3. Whilst thought to be mediated by autoimmune processes2, the exact pathogenesis of MS still remains unclear and at present there is no definitive understanding of the significance of epidemiological data or indeed the efficacy of potential prevention and treatment methods.
In 2008, Zamboni et. al published a paper that described compromised venous drainage in the central nervous system and hypothesised that this contributed to the development of MS, referring to this concept as chronic cerebrospinal venous insufficiency (CCSVI)4. They went on to describe a surgical technique, branded the ‘liberation procedure’ by the media, which attempted to improve compromised blood flow, via angioplasty5.
My neighbour, Sarah (48), was diagnosed with MS in 2004 and over the past decade the disease has had a major impact on her life, forcing her to work from home due to its debilitating effects on her movement and strength. When Zamboni et. al’s findings were first announced, Sarah, like thousands of other MS sufferers took great interest in the role of CCSVI in MS. She was hopeful that these new developments could one day improve her own condition, a hope that was augmented by online MS forums reporting positive effects of CCSVI treatments in other countries.
However, CCSVI has been met with some scepticism within the medical community. Studies have shown conflicting results, with some researchers supporting Zamboni et. al’s work, some reporting contrasting findings and others even refuting the existence of CCSVI.
NICE guidelines state that “current evidence on the efficacy of percutaneous venoplasty for CCSVI for MS is inadequate in quality and quantity. Therefore, this procedure should only be used in the context of research”6. However, experimental CCSVI procedures are reportedly being performed in other parts of the World, which is alarming as CCSVI intervention has been associated with serious injuries and even death7.
Understanding the relationship between CCSVI and MS is imperative, both in protecting patients from potentially harmful interventions, and providing patients like Sarah with accurate information about appropriate treatment options.
Literature Review
The following section of this essay will appraise relevant articles that have employed a variety of methods and techniques (outlined in Table 1) to examine the relationship between CCSVI and MS.
Table 1: Technique abbreviations
Technique | Abbreviation |
Transcranial colour-coded and echocolour Doppler sonography | TCCS-ECD |
Blood volume flow | BVF |
Cross-sectional area | CSA |
Colour-coded duplex sonography | CCD |
CCSVI in patients with MS4
This pioneering study by Zamboni et. al was the first to investigate extracranial venous outflow routes in clinically defined MS4. In this respect, the study could be thought to be hypothesis-generating. The authors give context to their study by referencing previous work, acknowledging that “MR venography and post-mortem studies have demonstrated a topographic correspondence between MS plaques and the cerebral venous system”.
65 patients with MS and 235 controls blindly underwent TCCS-ECD. Detection of at least two of the five parameters of venous outflow (Table 2) constituted CCSVI diagnosis. All the MS patients but none of the controls were positive for CCSVI. This 100% specifity, sensitivity and positive/negative predictive value could reflect spectrum bias.
Table 2: Five parameters of CCSVI.
1 | Reflux in the Internal Jugular Veins (IJV) and/or Vertebral veins (VV) in sitting and supine posture |
2 | Reflux in the Deep Cerebral Veins |
3 | High-resolution B-mode evidence of IJV stenoses |
4 | Flow not Doppler-detectable in the IJVs and/or VVs |
5 | Reverted postural control of the main cerebral venous outflow pathways |
Subsequently, all MS patients and 48 controls without neurological disease but scheduled for venography, underwent selective venography of the azygos and jugular venous system with venous pressure measurement. This revealed multiple severe extracranial stenosis in all 65 of the MS patients, but in none of the controls.
The authors highlight that performing the TCCS-ECD protocol once with a single team of investigators is a limitation that questions the reproducibility of the results. However, they fail to recognise that examining “at least one of the deep cerebral veins” introduces variation into the results as ideally all three of the deep veins should have been examined. This compromises the validity of the findings.
Despite declaring no competing interests, reports emerged after the study was published revealing that Zamboni’s wife had been diagnosed with MS8 and that he had financial ties to Esaote, the manufacturer of the ultrasound equipment specifically used in CCSVI diagnosis9. These vested interests are a potential source of bias.
Whilst the authors acknowledge that further research is required,they state that their TCCS-ECD protocol should be utilised when a patient first presents with demyelinating symptoms. This conclusion goes beyond these preliminary findings, assuming the results are proof of the pathogenic theory.
No Cerebrocervical Venous Congestion in Patients with MS10
After the previous paper was published there was a swift response from researchers attempting to replicate the results. One such study is from Doepp et. al which clearly challenges the CCSVI concept10.
A strength of this study is its use of additional ultrasound indices like BVF, CSA of IJV and IJV flow analysis. The author’s state that these additions will allow them to “more comprehensively evaluate the haemodynamic effects of any suspected cerebrocervical congestion”, contributing to the wider field of research. The authors criticise Zamboni et. al4 for not reporting BVFs in their findings. They point out that analysis of reflux using only the Doppler spectrum “may lead to misinterpretation of blood flow”. Indeed, a review by Laupacis et. al also questions the reliability of ultrasound assessment11.
41 RRMS patients, 15 SPMS patients and 20 controls were examined with TCCS-ECD and the additional indices but in contrast to the study by Zamboni et. al4, none of the subjects fulfilled more than one of the criteria for CCSVI. There was no detected IJV stenosis, no intracranial vein difference between groups and blood flow direction in the IJVs and VVs was normal in all subjects except for one patient. Compared to controls, a higher BVF was detected in patients with MS in an upright position, whereas in a lying position no difference emerged. Interestingly, the authors point out that “higher BVF in patients should suggest a better than normal cerebral venous drainage”, questioning the existence of CCSVI.
The study was performed in a single centre with a small cohort that was limited to only two sub-groups, making results less generalisable. Additionally, the investigators were not blinded, introducing a potential source of bias. The authors recognise these limitations yet insist they do not “significantly compromise the validity of our findings”, which is questionable.
Flow direction was assessed during the Valsalva manoeuvre whereas Zamboni et. al4 used physiological breathing. This alteration undermines the author’s aim of evaluating Zamboni et. al’s4 results. Furthermore, there is no evidence to confirm that the investigators had adequate experience to conduct the tests and interpret the results, reducing validity. Moreover, the small number of controls could lead to selection bias.
The authors call for more research in the subject and discourage clinical intervention.
Prevalence of extracranial venous narrowing on catheter venography in people with MS, their siblings and unrelated healthy controls12
This study by Traboulsee et. al has a clear primary aim, “to clarify whether venous blockages are unique to multiple sclerosis”12. Innovatively, it is the first study to validate its imaging techniques with catheter venography. The author’s state that this is regarded as the “gold standard for the assessment of venous stenosis”, giving “direct visualisation of venous anatomy”.
An assessor-blinded, case-control study was carried out over three different centres. These measures were appropriate for increasing reliability and making results more generalisable. The authors investigated over 50% narrowing of the internal jugular and azygos veins with catheter venography. This was combined with Doppler ultrasound utilising Zamboni’s criteria.
The study included 79 patients with MS, 55 siblings and 43 unrelated controls. Not all participants were able to undergo all procedures for various reasons which are clearly stated by the authors. 171 participants underwent ultrasound, 149 underwent catheter venography and 143 underwent both procedures. The results are summarised below in Figure 1.
Figure 1: Traboulsee et. al’s12 results
The high rate of venous narrowing in the control group had not previously been reported. These results suggest that venous narrowing is common in the general population and is not a unique anatomical feature associated with MS.
Another important finding was that CCSVI was rarely identified by catheter venography despite the high prevalence detected by ultrasound. This discrepancy has clinical relevance, implying that angioplasty should not be performed on the basis of single modality studies.
One strength of the study is the examination of unaffected siblings of patients, as susceptibility to MS has been shown to have a genetic link13. Additionally, the ultrasound equipment was identical to that used by Zamboni et. al4 and the ultra-sonographers were trained by Zamboni in Italy. These measures were appropriate in increasing the validity of the findings.
The authors declare financial sponsorship but insist that the sponsors had no role in study design. However, funding from the MS Society of Canada is of interest as Canada has one of the highest incidences of MS in the World and Canadian health bodies have been under immense pressure to make CCSVI treatment available to interested MS patients7.
Normal CSF ferritin levels in MS suggest against etiological role of CCSVI14
Singh and Zamboni have suggested that disease progression in MS may be related to iron deposition following CCSVI15. A prospective study by Worthington et. al examined this potential mechanism by investigating CSF ferritin levels in patients with MS14. The author’s hypothesise that “ongoing parenchymal iron deposition due to CCSVI should lead to an increase in CSF ferritin over time”. The study type was effective as prospective studies usually have fewer sources of bias than retrospective studies.
22 PPMS, 20 SPMS and 9 RRMS patients were recruited from a previously published Dutch cohort and a group of 1071 controls was established consisting of patients presenting with headaches with normal CT brain scans. Groups of patients with sub-arachnoid haemorrhage, superficial siderosis and meningoencephalitis were also included. CSF ferritin levels were measured following lumbar puncture and were defined as pathological if over 12 ng/mL. MRI scans were then performed to assess T1 and T2 lesion volume.
The cross-sectional part of the study found that pathologic CSF ferritin levels were observed in 10% of RRMS patients, 11% of PPMS patients, 23% of SPMS patients and 4% of controls. Higher levels were recorded in the other conditions. There was no positive correlation between CSF ferritin levels and tumour volume in patients with MS. In fact, patients with SPMS showed an inverse correlation between CSF ferritin levels and T2 lesion volume.
The longitudinal part of the study examined 29 of the patients with MS after a three year period. This was appropriate for analysing the correlation between CSF ferritin levels and disease progression. Except for one patient with SPMS, CSF ferritin did not increase and the four patients with elevated levels at baseline had normalised at follow-up, contradicting the notion of a gradual build-up of iron. The relative increase of CSF ferritin in the SPMS patient was associated with improvement of lower limb function and reduction of T1 lesion volume. Following this “paradoxical finding”, the authors sensibly suggest that future research should focus on CSF ferritin level increase over time in MS.
The authors note that elevated CSF ferritin is not specifically indicative of iron deposition, but can also be observed with inflammation, a key pathological feature in MS1. They also point out that due to “ethical reasons”, they are unable to provide longitudinal data on CSF ferritin levels from control patients. This reduces the validity of the longitudinal results.
Despite the large control group, the small number of MS patients examined makes results less generalisable. Furthermore, it is not clear whether patients were undergoing treatment at the time of ferritin analysis. Indeed, a separate study reported that serum ferritin levels were significantly increased in patients with RRMS at 12 months after initiating interferon-B therapy16. This may have clinical relevance as CSF ferritin levels could have a potential role in monitoring an individual’s response to interferon-B therapy.
MS and CCSVI: A Population-Based Case Control Study17
This study by Patti et. al was the first to adopt a population-based control design17. Whilst appropriate for identifying association, this type of study can be prone to bias when selecting participants18. However, the authors employed rigorous measures to overcome this issue. An MS incident cohort was established using epidemiological surveys carried out between 1975 and 2004 in Catania, Sicily. Following this, a computer program randomly selected individuals to be invited to participate in the study, removing selection bias and making results more generalisable. Multistage sampling methods were employed to establish a control group from the general population of Catania that was matched by age and gender. Extensive exclusion criteria was applied to both patients and controls to remove confounding factors such as pre-existing neurological and vascular conditions. This increased validity of results.
The study examined 148 MS patients, 20 patients with clinically isolated syndrome (CIS), 40 with other neurological disorders (OND) and 172 healthy controls. TCCS-ECD was performed by a single experienced vascular sonographer who had been trained by Zamboni.
CCSVI was present in 18.9% of MS patients, 10% of CIS patients, 5% of OND patients and 6.4% of controls. An association was found between MS and CCSVI with an odds ratio of 3.41. CCSVI was also significantly more frequent in MS patients with longer disease duration and in those with the progressive forms of the disease. Consequently, the authors draw the conclusion that “CCSVI could be related to MS disability”.
The authors highlight that despite finding a positive correlation between CCSVI and MS, the results are still not as strong as those reported by Zamboni et. al4. They suggest that this variability may be related to the accuracy and reliability of ultrasound. This, coupled with the use of a single sonographer, questions the reproducibility of results.
The author’s reference previous work, acknowledging that “poor reporting of the success of blinding” has limited the definite conclusions that can be made from studies that have found a positive association between CCSVI and MS. In this respect, the study enhances existing research by blinding both the sonographer and adjudication panel to subject’s status.
Traboulsee et. al comment on the “highly dynamic and variable nature of the venous system”12. This concept is also eluded to by the authors of this paper, who call for future studies investigating healthy subjects at different ages to help understand the “physiological variability” of the venous system.
No Evidence of CCSVI at MS onset19
Baracchini et. al hypothesise that “if a cause-effect relationship between CCSVI and MS exists, this should be observed at disease onset”19. This study adds to the literature as previous research has not focused on CCSVI at MS onset. The authors give context to their study by providing a comprehensive background to the CCSVI debate.
A strength of the study is the rigorous diagnostic mark-up which included CSF analysis, brain and spinal cord MRI imaging, blood tests and echo-colour Doppler sonography. Evidence of dissemination in space of inflammatory lesions was used to determine if a CIS was possibly MS (pMS). Over one year, 50 consecutive patients presenting with pMS were recruited and underwent the diagnostic workup. Those with CCSVI then underwent selective venography.
60 patients with transient global amnesia (TGA), a condition known to occur in association with venous anomalies, were also studied. For both pMS and TGA groups there were identical sized healthy control groups that were age and gender matched, increasing validity. These three groups also underwent sonography.
Transcranial sonography findings were normal in all the pMS patients. One or more abnormal extracranial sonography findings were observed in 52.0% of pMS patients, 31.8% of healthy controls and 68.3% of TGA patients. 8 pMS patients fulfilled the diagnosis of CCSVI and in the 7 assessed (one denied consent), no venous anomalies were found.
As seen in the study by Traboulsee et. al12, there was a discrepancy between the two modalities, with the sonography showing an association between CCSVI and MS that was not present in venography. The authors highlight that analysis of BVF, as seen in the study by Doepp et. al10, may explain these differences. Unfortunately, the authors fail to document any internal validation of their methods, making it unclear whether the sonography data is more or less valid than the venography data.
The study is described as blinded but there is no indication of how the blinding was ensured or whether it was tested. Furthermore, no measures were taken to blind the radiologist from the clinical diagnosis of the patients undergoing selective venography, introducing a source of bias.
The cohort had a lower mean age compared to that studied by Zamboni et. al4. This may remove some bias as a previous large study of healthy volunteers showed that retrograde venous flow was significantly more frequent in those over fifty20.
The authors suggest that future studies should aim to “elucidate whether MS-associated pathology may contribute to determine a CCSVI condition”. They highlight that the high percentage of venous abnormalities seen in the non-MS groups are in line with previous work and emphasise that invasive therapeutic procedures are “not only dangerous but presently unjustified”.
CCSVI in MS: the CoSMo study21
Comi et. al performed an observational, case-control study involving 35 centres across Italy. The study contributes to the research field as it is the largest to date, making results more generalisable.
The prevalence of CCSVI was examined in 1165 patients with MS, 226 with OND and 376 healthy controls. CCSVI assessment using the Zamboni criteria was performed following CCD examination. Specific training and a final examination were required for each sonologist. This increased reliability and addressed possible limitations due to poor training. After the sonologist had performed their blinded investigation and made their diagnosis, video clips of the CCD examination were sent at random to one of 3 central expert sonologists whom performed a second blinded reading. This design was effective in reducing bias and increasing reliability.
CCSVI prevalence was 3.26%, 3.10% and 2.13% for the MS, OND and HC groups respectively. Additionally, MS patient sub-group analysis did not reveal an association with CCSVI and disease course, contradicting results from Patti et. al17. Although the PPMS group had a slightly higher CCSVI prevalence, it did not attain statistical significance.
The difference in the prevalence of CCSVI between the participating centres was high, with some centres having a prevalence of 50-60% and others zero. The authors suggest that this is a result of the tendency of the local examiner to “see” the abnormalities. This idea is supported by the fact that regardless of this propensity there was never a difference in CCSVI prevalence among the three groups. This highlights the importance of blinding to avoid biased results.
There was good agreement between the local and central examiners with regard to the absence of CCSVI but there was a poor agreement in the positive diagnosis of CCSVI. The authors speculate that this is due to problems with the CCD examination but also suggest that the Zamboni criteria may contribute.This suggests that the Zamboni criteria is too subjective to interpretation and should be reviewed to minimise bias.
The authors performed a power calculation to calculate an appropriate sample size. This rigorous measure is beneficial as underpowered studies have been associated with false negative results18. The authors also highlight that effective controls helped account for variability, with the multi-centric design and large sample size partially controlling for instrument variation and the complete blinding of the central reader controlling for “inter-user variability and subjectivity at the local level”. However, it is not indicated whether the control group was matched for age and gender.
Whilst the authors reach the conclusion that CCSVI is not associated with MS, they recognise that “recruitment of unimpaired or minimally impaired patients” may account for the differences observed between their study and others.
Common limitations
Perhaps the most important limitation to be highlighted is that none of the studies are able to prove a temporal relationship between CCSVI and MS. Additionally, there is no standardised criteria used to define MS or MS severity which makes the studies less comparable. Furthermore, with the exception of Worthington et. al14, all the studies discussed tested for CCSVI using the Zamboni criteria, which has been criticised for both conceptual and technical reasons22, 23.
Conclusion
Zamboni et. al’s initial reports prompted hastened attempts by researchers to confirm a link between CCSVI and MS. The pressure to produce results led to many small, monocentric studies with various methodological flaws, whose conflicting results have only served to make the role of CCSVI all the more ambiguous. Indeed, two independent systematic reviews were unable to make definite conclusions about a potential association between CCSVI and MS24, 25.
The evidence presented in this essay shows that it is difficult to ascertain whether CCSVI is causal in the pathogenesis of MS or a side effect of the condition. It is also not clear whether CCSVI determines disease development. The majority of the literature points towards CCSVI not being an aetiological or influencing factor in MS but whilst there is still ambiguity surrounding the subject, further research is needed to both protect and provide realistic expectations for patients like Sarah.
Research Question
“Is CCSVI present at MS onset and does venous congestion alter during disease progression?”
Justification
Several studies have investigated CCSVI prevalence in patient cohorts, however multi-modal imaging techniques have not been used to study venous congestion at MS onset and throughout disease progression.
Building on Baracchini et. al’s19work, which examined patients at onset, a longitudinal cohort study could give a more reliable indication of the role of CCSVI in MS over time.
Whilst it is difficult to prove causality, examining a range of patients at different disease stages over an extended period may indicate whether CCSVI is an influencing factor in MS progression. This could potentially justify closer monitoring of venous congestion in patients to predict disease course and hence determine appropriate treatment.
Methodology
A power calculation will be performed to ascertain an appropriate sample size. Participants will then be randomly selected over multiple centres from the following groups:
- pMS
- RRMS
- PPMS
- SPMS
- OND
The groups will be matched for age and gender with healthy controls to increase validity.
Any participants with an existing condition or undergoing treatment that could interfere with the objective of the study will be excluded.
Specifically trained technicians will perform a diagnosis involving TCCS-ECD. These results will be validated by catheter venography, BVF and MRI. Standardised criteria specific to these imaging techniques would need to be ascertained and validated beforehand, although this is beyond the scope of this essay. The current CCSVI criteria would also need to be revised in relation to flaws identified by other researchers22, 23.
Investigators will be fully blinded to subject’s status and contact time will be kept to an absolute minimum to reduce bias. The examination will be video recorded and a panel of independent trained invigilators will also report their diagnosis for comparison to increase reliability.
Participants will be followed up after a three and six year period and results compared to baseline.
Word Count
Words in main body of text | 3882 |
Words in figures and tables | 117 |
Total words | 3999 |
List of Tables:
Table 1: Technique abbreviations
Table 2: Five parameters of CCSVI. These are absent in normal subjects.
List of Figures:
Figure 1: Results from the study by Traboulsee et. al
References:
1 Kumar PJ, Clark ML. Kumar & Clark’s clinical medicine. 7th ed. Edinburgh ; New York: Elsevier/Saunders 2009.
2 Poser CM. An atlas of multiple sclerosis. New York: Parthenon Pub. Group 1998.
3 Warren S, Warren KG, World Health Organization. Multiple sclerosis. Geneva: World Health Organization 2001.
4 Zamboni P, Galeotti R, Menegatti E, Malagoni AM, Tacconi G, Dall’Ara S, et al. Chronic cerebrospinal venous insufficiency in patients with multiple sclerosis. Journal of Neurology, Neurosurgery & Psychiatry. 2009;80(4):392-9.
5 Zamboni P, Galeotti R, Menegatti E, Malagoni AM, Gianesini S, Bartolomei I, et al. A prospective open-label study of endovascular treatment of chronic cerebrospinal venous insufficiency.[Erratum appears in J Vasc Surg. 2010 Apr;51(4):1079]. Journal of Vascular Surgery. 2009;50(6):1348-58.e1-3.
6 National Institute for Health and Clinical Excellence. Percutaneous venoplasty for chronic cerebrospinal venous insufficiency for multiple sclerosis: NICE; 2012.
7 Snyder J, Adams K, Valorik CA, Whitehurst D, Vallee J. “I knew what was going to happen if I did nothing and so I was going to do something”: Faith, hope, and trust in the decisions of Canadians with multiple sclerosis to seek unproven interventions abroad. BMC Health Services Research,. 2014;14(445).
8 Pullman D, Zarzeczny A, Picard A. “Media, politics and science policy: MS and evidence from the CCSVI Trenches”. BMC Medical Ethics. 2013;14:6.
9 Stone K. Medical device conflict of interest in the CCSVI debate. Annals of Neurology. 2012;71(3):A6-8.
10 Doepp F, Paul F, Valdueza JM, Schmierer K, Schreiber SJ. No cerebrocervical venous congestion in patients with multiple sclerosis. Annals of Neurology. 2010;68(2):173-83.
11 Laupacis A, Lillie E, Dueck A, Straus S, Perrier L, Burton JM, et al. Association between chronic cerebrospinal venous insufficiency and multiple sclerosis: a meta-analysis. CMAJ Canadian Medical Association Journal. 2011;183(16):E1203-12.
12 Traboulsee AL, Knox KB, Machan L, Zhao Y, Yee I, Rauscher A, et al. Prevalence of extracranial venous narrowing on catheter venography in people with multiple sclerosis, their siblings, and unrelated healthy controls: a blinded, case-control study. Lancet. 2014;383(9912):138-45.
13 Dyment DA, Ebers GC, Sadovnick AD. Genetics of multiple sclerosis. Lancet Neurology. 2004;3(2):104-10.
14 Worthington V, Killestein J, Eikelenboom MJ, Teunissen CE, Barkhof F, Polman CH, et al. Normal CSF ferritin levels in MS suggest against etiologic role of chronic venous insufficiency. Neurology. 2010;75(18):1617-22.
15 Singh AV, Zamboni P. Anomalous venous blood flow and iron deposition in multiple sclerosis. Journal of Cerebral Blood Flow & Metabolism. 2009;29(12):1867-78.
16 Sena A, Pedrosa R, Ferret-Sena V, Cascais MJ, Roque R, Araujo C, et al. Interferon beta therapy increases serum ferritin levels in patients with relapsing-remitting multiple sclerosis. Multiple Sclerosis. 2008;14(6):857-9.
17 Patti F, Nicoletti A, Leone C, Messina S, D’Amico E, Lo Fermo S, et al. Multiple sclerosis and CCSVI: a population-based case control study. PLoS ONE [Electronic Resource]. 2012;7(8):e41227.
18 Greenhalgh T. How to read a paper : the basics of evidence-based medicine. 4th ed. Chichester, West Sussex, UK ; Hoboken, NJ: Wiley-Blackwell 2010.
19 Baracchini C, Perini P, Calabrese M, Causin F, Rinaldi F, Gallo P. No evidence of chronic cerebrospinal venous insufficiency at multiple sclerosis onset. Annals of Neurology. 2011;69(1):90-9.
20 Akkawi NM, Agosti C, Borroni B, Rozzini L, Magoni M, Vignolo LA, et al. Jugular valve incompetence: a study using air contrast ultrasonography on a general population. Journal of Ultrasound in Medicine. 2002;21(7):747-51.
21 Comi G, Battaglia MA, Bertolotto A, Del Sette M, Ghezzi A, Malferrari G, et al. Observational case-control study of the prevalence of chronic cerebrospinal venous insufficiency in multiple sclerosis: results from the CoSMo study. Multiple Sclerosis. 2013;19(11):1508-17.
22 Mayer CA, Pfeilschifter W, Lorenz MW, Nedelmann M, Bechmann I, Steinmetz H, et al. The perfect crime? CCSVI not leaving a trace in MS. Journal of Neurology, Neurosurgery & Psychiatry. 2011;82(4):436-40.
23 Costello F, Modi J, Lautner D, Bhayana D, Scott JN, Davenport WJ, et al. Validity of the diagnostic criteria for chronic cerebrospinal venous insufficiency and association with multiple sclerosis. CMAJ Canadian Medical Association Journal. 2014;186(11):E418-26.
24 Thapar A, Lane T, Nicholas R, Friede T, Ellis M, Assenheim J, et al. Systematic review of sonographic chronic cerebrospinal venous insufficiency findings in multiple sclerosis. Phlebology. 2011;26(8):319-25.
25 Zaniewski M, Kostecki J, Kuczmik W, Ziaja D, Opala G, Swiat M, et al. Neck duplex Doppler ultrasound evaluation for assessing chronic cerebrospinal venous insufficiency in multiple sclerosis patients. Phlebology. 2013;28(1):24-31.