Anthony L Traboulsee, Katherine B Knox, Lindsay Machan, Yinshan Zhao, Irene Yee, Alexander Rauscher, Darren Klass, Peter Szkup, Robert Otani, David Kopriva, Shanti Lala, David K Li, Dessa Sadovnick
Lancet 2014; 383: 138–45
Published Online October 9, 2013http://dx.doi.org/10.1016/ S0140-6736(13)61747-X
Department of Medicine (A L Traboulsee MD, Y Zhao PhD, Prof D Sadovnick PhD), Department of Radiology (L Machan MD, A Rauscher PhD, D Klass MD, Prof D K Li MD), and Department of Medical Genetics (I Yee MSc, D Sadovnick), University of British Columbia, Vancouver, BC, Canada; Department of Physical Medicine and Rehabilitation (K B Knox MD) and Department of Medical Imaging (P Szkup MD, R Otani MD), College of Medicine, University of Saskatchewan, Saskatoon, Canada; and Department of Surgery, University of Saskatchewan, Regina, Saskatchewan, Canada (D Kopriva MD, S Lala MD)
Correspondence to: Dr Anthony L Traboulsee, University of British Columbia Hospital, University of British Columbia, 2211 Wesbrook Mall, Room s199, Vancouver, BC, V6T 2B5, Canada firstname.lastname@example.org
Background Chronic cerebrospinal venous insuﬃciency has been proposed as a unique combination of extracranial venous blockages and haemodynamic ﬂow abnormalities that occurs only in patients with multiple sclerosis and not in healthy people. Initial reports indicated that all patients with multiple sclerosis had chronic cerebrospinal venous insuﬃciency. We aimed to establish the prevalence of venous narrowing in people with multiple sclerosis, unaﬀected full siblings, and unrelated healthy volunteers.
Methods We did an assessor-blinded, case-control, multicentre study of people with multiple sclerosis, unaﬀected siblings, and unrelated healthy volunteers. We enrolled the study participants between January, 2011 and March, 2012, and they comprised 177 adults: 79 with multiple sclerosis, 55 siblings, and 43 unrelated controls, from three centres in Canada. We assessed narrowing of the internal jugular and azygous veins with catheter venography and ultrasound criteria for chronic cerebrospinal venous insuﬃciency proposed by Zamboni and colleagues. Catheter venography data were available for 149 participants and ultrasound data for 171 participants.
Findings Catheter venography criteria for chronic cerebrospinal venous insuﬃciency were positive for one of 65 (2%) people with multiple sclerosis, one of 46 (2%) siblings, and one of 32 (3%) unrelated controls (p=1·0 for all comparisons). Greater than 50% narrowing of any major vein was present in 48 of 65 (74%) people with multi- ple sclerosis, 31 of 47 (66%) siblings (p=0·41 for comparison with patients with multiple sclerosis), and 26 of 37 (70%) unrelated controls (p=0·82). The ultrasound criteria for chronic cerebrospinal venous insuﬃciency were fulﬁlled in 35 of 79 (44%) participants with multiple sclerosis, 17 of 54 (31%) siblings (p=0·15 for comparison with patients with multiple sclerosis) and 17 of 38 (45%) unrelated controls (p=0·98). The sensitivity of the ultrasound criteria for detection of greater than 50% narrowing on catheter venography was 0·406 (95% CI 0·311–0·508), and speciﬁcity was 0·643 (0·480–0·780).
Interpretation This study shows that chronic cerebrospinal venous insuﬃciency occurs rarely in both patients with multiple sclerosis and in healthy people. Extracranial venous narrowing of greater than 50% is a frequent ﬁnding in patients with multiple sclerosis, unaﬀected siblings, and unrelated controls. The ultrasound criteria are neither sensitive nor speciﬁc for narrowing on catheter venography. The signiﬁcance of venous narrowing to multiple sclerosis symptomatology remains unknown.
Funding MS Society of Canada, Saskatoon City Hospital Foundation, Lotte and John Hecht Memorial Foundation, Vancouver Coastal Health Foundation, and the Wolridge Foundation.
Multiple sclerosis is a leading cause of neurological disability and is estimated to aﬀect more than 2 million people worldwide. The cause of the disease remains uncertain and a vascular mechanism has long been suggested to have a possible role. Zamboni and colleagues reported that the presence of multiple stenosis of the extracranial venous drainage system—a disorder they called chronic cerebrospinal venous insuﬃciency— was signiﬁcantly associated with multiple sclerosis. The venous blockages were present in all 65 patients with multiple sclerosis studied by ultrasound and on catheter venography. These combinations of blockages were not seen in healthy control participants studied by ultrasound, nor in patients with other diseases who underwent catheter venography. Zamboni and colleagues thus speculated that venous blockages have a central role in the pathogenesis of multiple sclerosis, and that treatment with venoplasty would ameliorate the disease.
An unblinded, uncontrolled interventional treatment trial by Zamboni and colleagues5 of venoplasty in patients with multiple sclerosis showed improvements in disability and quality of life. The procedure was hailed by media outlets as a “medical miracle”, and many patients with multiple sclerosis advocated for treatment before independent validation studies had been completed. Thousands of venoplasties, venous stenting procedures, or both, were undertaken outside registered clinical trials, despite expressed safety concerns, contradictory results from imaging studies, and caution recommended by the Society for Interventional Radiology, patient advocacy societies, the US Food and Drug Administration, and the UK National Institute for Health and Care Excellence.
The inability of independent research groups to reproduce Zamboni and colleagues’ ﬁndings regarding the diagnosis of chronic cerebrospinal venous insuﬃciency in all patients with multiple sclerosis called into question the existence of the disorder and its role in multiple sclerosis pathogenesis. However, until now, studies have been limited to the imaging techniques of ultra- sound and magnetic resonance venography without validation from catheter venography, which has the advantage of direct visualisation of venous anatomy. Catheter venography is regarded as the gold standard for the assessment of venous stenosis because it details the venous anatomy with high spatial and temporal resolution and allows imaging of the azygous vein. Collateral ﬂow, stasis, and reﬂux patterns can be assessed in real time. Changes in cerebral venous drainage with posture are profound.20 In the supine position, venous ﬂow occurs mainly through the jugular veins, whereas in the upright position the ﬂow is redirected through the vertebral venous plexus and azygous vein. Changes in intrathoracic or intra-abdominal pressure (eg, Valsalva manoeuvre) also aﬀect drainage. In this Article, we present a blinded study detailing venous anatomy with catheter venography in patients with multiple sclerosis, their unaﬀected siblings who are at an increased risk of developing the disease, and age-matched and sex- matched unrelated healthy controls. Genetic factors aﬀect susceptibility to multiple sclerosis21 and clinical course. The risk of multiple sclerosis can be assessed according to the degree of genetic sharing with an aﬀected patient. The overall lifetime risk of multiple sclerosis in the white population is two in 1000, 35 in 1000 for a sibling of a person with multiple local newspapers. Inclusion criteria for patients with multiple sclerosis were age 18–65 years, diagnosis of deﬁnite multiple sclerosis (2010 McDonald criteria), and an expanded disability status score29 between 0 and 6·5. We excluded patients who had previously undergone venous procedures or those using vasoactive drugs.
All assessing study team members were masked to participant status (patient with multiple sclerosis, un- aﬀected sibling, or unrelated control). Participants were positioned and draped (covered with a blanket, leaving only the head and neck exposed) by the unmasked study coordinator, mobility aids were hidden, and participants were instructed to not discuss their diagnosis with the study team. Blinding was assessed by questionnaire, which asked the study team members whether or not they thought each participant had multiple sclerosis. Clinical assessments included expanded disability status score29 and relapses within the previous 2 years.
The study protocol was approved by clinical research ethics boards at the three participating Canadian centres (University of British Columbia Hospital, Vancouver; University of Saskatchewan, Saskatoon; and Regina Qu’Appelle Health Region, Regina, Saskatchewan). All participants provided written informed consent.
In view of the plastic nature of veins with postural changes, we did catheter venography on a tilt table in the supine position, with the participants positioned 85° upright to assess venous anatomy in the best possible way. We standardised the protocol across the three centres, with three training sessions for the performance and interpretation of venography. We introduced a sheath into the right femoral vein in the supine position with use of local anaesthetic and ultrasound guidance. We advanced a 5 Fr hockey stick or H1 catheter over a guidewire under ﬂuoroscopic guidance into the right internal jugular vein at the level of the skull base, 5 cm below the mastoid, with the tip directed laterally. We tilted the table upright and, after three in-and-out respirations to enhance change in positional cranial venous drainage, we recorded intravenous pressure. We did pump injections at 2 cm³/s to assess reﬂux and 4 cm³/s to assess anatomy during quiet apnoea and normal exhalation. We returned the table to the supine position and, after three in-and-out respirations, repeated the same procedures with the catheter in the upper and lower jugular vein. We repeated all procedures in the left internal jugular vein and the azygous vein. We recorded the upper, lower, and withdrawal pressures for the azygous vein relative to the superior vena cava. We deﬁned internal jugular vein stenosis as a narrowing of more than 50% (valvular or non-valvular vessel segment) in comparison ith a normal reference segment (widest vessel segment below the mandible) in the supine position. We used the supine position for this assessment because both internal jugular veins usually collapse in the upright position, with blood redirected to the azygous vein and vertebral veins. We deﬁned azygous vein stenosis as more than 50% narrowing relative to the largest normal segment in either the upright or supine position. We judged a haemodynamically signiﬁcant narrowing to be present if at least one of the following was recorded: reﬂux (persistent retrograde ﬂow of most of the contrast bolus after injection was completed); stasis (contrast was present after the injection); or abnormal collaterals (one or more vessels >50% the size of the adjacent primary vessel, or two or more collateral vessels present at <50% the size of the adjacent primary vessel). We assessed the four patterns of venous obstruction described by Zamboni and colleagues, including stenosis of the proximal azygous vein and complete occlusion of one internal jugular vein (type A); stenosis of both internal jugular veins and the proximal azygous vein (type B); bilateral stenosis of internal jugular veins only (type C); and azygous vein stenosis (type D).
Venous doppler ultrasound equipment was identical to that used in Zamboni and colleagues’ study4 (MyLab Vinco, Esaote, Genoa, Italy [Health Canada, Health Products and Food Branch medical device #79702]). Experienced ultrasonographers were trained by Zamboni in Ferrara, Italy. We used a 7 MHz transducer capable of grey scale (for venous stenosis and loss of postural control of venous outﬂow) and duplex (for reﬂux) to study the internal jugular vein from the skull base to as far centrally as was visible.
Table 1: Baseline demographics and clinical characteristics
Participants with multiple sclerosis Healthy controls (n=98)
Female 67 (85%) 71 (72%)
Male 12 (15%) 27 (28%)
British Columbia 42 (53%) 63 (64%)
Saskatchewan 37 (47%) 35 (36%)
Mean (SD) 47·8 (10·1) 48·7 (10·7)
Median (range) 48·0 (22·0–68·0) 49·0 (27·0–72·0)
Expanded disability status score
Mean (SD) 2·7 (1·6) NA
Median (range) 2·0 (0–6·5) NA
Multiple sclerosis duration, years
Mean (SD) 16·1 (9·3) NA
Median (range) 13·0 (2·0–41·0) NA
Disease course, n
Relapsing–remitting 67 (85%) NA
Secondary progressive 8 (10%) NA
Primary progressive 4 (5%) NA
Data are n (%), unless otherwise indicated. NA=not applicable.
…cont’d from above
We measured the internal jugular vein cross-sectional area in both the supine and upright positions during a short period of apnoea after a normal exhalation. A cross-sectional area smaller than 0·3 cm² was judged to be a positive result. Loss of postural control of cerebral venous outﬂow was established by subtraction of the cross-sectional area of the internal jugular vein in the upright position from that in the supine position measured during a short period of apnoea after a normal exhalation. Normal response is an increased diameter of the internal jugular vein in the supine position compared with the upright position. Normal direction of ﬂow is towards the chest in both positions recorded during a short period of apnoea after exhalation. Reﬂux was present when a cranially directed ﬂow of 0·88 s or longer was identiﬁed. We recorded the direction of ﬂow in the internal cerebral vein, basal vein of Rosenthal, and great vein of Galen in quiet respiration in both positions with a 2 MHz transcranial Doppler transducer. Reversal of ﬂow for longer than 0·5 s was viewed as abnormal. At least two positive ultrasound criteria were diagnostic for chronic cerebrospinal venous insuﬃciency.4
Data forms were reviewed and double entered into an electronic database at the University of British Columbia Coordinating Centre. Missing data ﬁelds or logical queries corrections occurred before ﬁnal database lock (on Oct 29, 2012) and statistical analyses. In the analyses, Sonographers correctly identiﬁed multiple sclerosis status in 37 of these 82 cases (45%) and responded with “no idea” in 16 (20%). Blinding was assessed after catheter venography in 78 of 149 (52%) cases. Radiologists selected “no idea” regarding multiple sclerosis status in all cases.
With Zamboni and colleagues’ catheter venography classiﬁcation of chronic cerebrospinal venous insufﬁciency,4 we identiﬁed one of 65 (2%) patients with multiple sclerosis, one of 46 (2%) unaﬀected siblings, and one of 32 (3%) unrelated controls who met the criteria (table 2). 48 of 65 (74%) people with multiple sclerosis, 31 of 47 (66%) siblings, and 26 of 37 (70%) unrelated controls met the criterion of more than 50% narrowing on venography in any major extracranial vein. When evidence of physiological ﬂow alterations was added to narrowing, 33 of 65 (51%) people with multiple sclerosis, 21 of 47 (45%) siblings, and 20 of 37 (54%) unrelated controls met the criteria. On the basis of the multivariate logistic regression, the odds of participants
Participants with multiple sclerosis
48/65 (74%) 20/27 (74%) 28/38 (74%) 1·06 (0·35–3·19) 0·92
Participants without multiple sclerosis
53/78 (68%) 21/29 (72%) 32/49 (65%) 1·50 (0·50–4·46) 0·66
meeting the criterion of more than 50% narrowing were higher at University of British Columbia Hospital than in Saskatchewan (OR 2·78, 95% CI 1·35–5·88, p=0·006). We recorded the same trend in patients with multiple laboratories, four exclusionary medications, two cases of orthostatic hypotension, one previous vascular procedure, one claustrophobia, one history of head injury, and one contrast allergy) and 17 people withdrew before either procedure was done. Of the 177 participants included, 171 had ultrasound, 149 catheter venography, and 143 both procedures. The reason for missing an ultrasound (six healthy controls) was travel distance (inability to schedule tests on the same day). Reasons for incomplete catheter venography studies (n=12) were the inability to tolerate the full protocol because of vasovagal symptoms or dis- comfort; one patient had an asymptomatic venous dissection. Reasons for missing catheter venography studies (n=16) included change in blood work (liver or renal function) and participant choice.
79 patients with multiple sclerosis, 55 unaﬀected siblings, and 43 unrelated healthy controls participated in the study, with a similar mean age and proportion of women in each group (table 1). The participants included 12 twin pairs (three monozygotic and nine dizygotic), one pair from a triplet, and 30 additional sibling pairs. Of these pairs, 19 were HLA identical. Two full sets of triplets also took part, one of which was HLA identical. 30 patients with multiple sclerosis without a sibling participated.
67 (85%) of the patients with multiple sclerosis had relapsing–remitting disease. The median expanded disability status score was 2·0 (range 0–6·5) and median disease duration was 13·0 years (2·0–41·0 years). Blinding was assessed after ultrasound in 82 of 171 cases. We recorded similar results with use of the catheter venography criteria of greater than 50% stenosis with evidence of ﬂow abnormalities with sensitivity 0·400 (95% CI 0·287–0·524), speciﬁcity 0·616 (0·495–0·726), positive predictive value 0·500 (0·373–0·627), and negative predictive value 0·517 (0·408–0·625). When we considered sibling pairs only, we noted no signiﬁcant diﬀerence in rates of participants who met ultrasound criteria (40 pairs) for chronic cerebrospinal venous insuﬃciency or any of the catheter criteria for narrowing (27 pairs; data not shown).
We present the ﬁrst study to use catheter venography systematically in patients with multiple sclerosis, their unaﬀected siblings, and unrelated healthy volunteers, with careful attention to standardisation of imaging protocols and blinding (panel). The results provide no evidence that extracranial venous anatomy diﬀers between patients with multiple sclerosis, their unaﬀected siblings (including monozygotic twin pairs), and unrelated healthy controls. Catheter venography ﬁndings of chronic cerebrospinal venous insuﬃciency as described by Zamboni and colleagues are therefore rare ﬁndings and do not distinguish patients with multiple sclerosis from healthy controls. Almost 75% of the total study sample had venous narrowing of more than 50% on catheter venography, with little diﬀerence recorded between patients with multiple sclerosis and controls. This high rate of narrowing in a healthy control group has not previously been reported. Moreover, with ultrasound, we did not ﬁnd diﬀerences in the rates of Zamboni-deﬁned chronic cerebrospinal venous insuﬃciency (two or more positive criteria of ﬁve) or valvular abnormalities between healthy controls and participants with multiple sclerosis. A meta-analysis of published studies of ultrasound diagnosis of chronic cerebrospinal venous insuﬃciency suggests an association between this disorder and multiple sclerosis (OR 3·5, 95% CI 0·8–15·8). In our study, the agreement between ultrasound and catheter venography for the detection of venous narrowing was very poor. Moreover, chronic cerebrospinal venous insuﬃciency was rarely detected by catheter venography despite the higher prevalence of cases meeting criteria for chronic cerebrospinal venous insuﬃciency on ultrasound. Our results also challenge both the validity of ultrasound for the purpose of detecting chronic cerebrospinal venous insuﬃciency and its existence as a disorder. Few studies of chronic cerebrospinal venous insufﬁciency in multiple sclerosis have included the gold standard method of catheter venography, despite the fact that this procedure has been proposed as a guide for decision making in endovascular treatment. Scalise and colleagues did a small study in patients with multiple sclerosis to show the advantages of intravascular ultrasound. In a smaller study of ten patients with multi- ple sclerosis, Zivadinov and colleagues compared magnetic resonance venography techniques with catheter venography. Both studies did not have controls or blinding, and they did not address the original descriptions of chronic cerebrospinal venous insuﬃciency.
Standardised testing protocols, blinding, and the inclusion of controls are essential for the scientiﬁc exploration of the theory of chronic cerebrospinal venous insuﬃciency. The highly dynamic and variable nature of venous anatomy in particular calls for a rigorous study design. The interobserver variability noted between our participating sites shows how crucial blinding is for studies of venous anatomy in which technical variability can occur. The subjective interpretation by unmasked observers of venography data could lead to biased results. Although we noted diﬀerences in the prevalence of narrowing between sites, these diﬀerences were recorded equally across all study groups, which supports the maintenance of blinding. The hypothesis that venous narrowings have a role in the cause of multiple sclerosis is unlikely, since the prevalence of venous narrowings is similar in people with the disease, unaﬀected siblings, and unrelated healthy controls on catheter venography. Narrowing of more than 50% was recorded in almost 75% of the study population (cases and controls) on catheter venography, which supports the contention that venous narrowing is a common anatomical variant. The signiﬁcance of venous narrowing to multiple sclerosis symptomatology remains unknown.
Our results conﬁrm that venous narrowing is a frequent ﬁnding in the general population and is not a unique anatomical feature associated with multiple sclerosis. We show that the proposed disorder of chronic cerebrospinal venous insuﬃciency deﬁned by ultrasound criteria is an insensitive screening test for venous narrowing present on catheter venography. Our results call into question the basis for the term chronic cerebrospinal venous insuﬃciency and its putative role in the pathogenesis of multiple sclerosis.
Panel: Research in context
We searched PubMed with combinations of the keywords “ccsvi”, “multiple sclerosis”, healthy controls”, “validation”, “ultrasound”, and/or “catheter venography”, and identiﬁed four original research articles that included results of both ultrasound or magnetic resonance venography and catheter venography. None of these reports were blinded studies that compared catheter venography results in patients with multiple sclerosis versus healthy controls. Three articles showed poor agreement with ultrasound and catheter venography, and a fourth reported poor agreement with magnetic resonance venography and catheter venography. None of the studies included a control group undergoing catheter venography.
ALT was the principal investigator of the study, and KBK (Saskatoon) and Fog T. On the vessel–plaque relations in brain in multiple sclerosis. Acta Neurol Scand 1963; 39: 258–62. DKo (Regina) were the site-speciﬁc principal investigators. ALT wrote the ﬁrst draft. All authors reviewed the report, and LM, KBK, DKL, DS, YZ, and IY provided input toward the ﬁnal version, which was prepared by ALT. ALT, LM, KBK, DKL, DS, YZ, IY, and PS designed the study. LM, SL, PS, RO, and DKL developed the procedures. RO reﬁned the study hypoperfusion in a central nervous system where mechanisms allowing leukocyte inﬁltration are readily upregulated?
Med Hypotheses 1998; 51: 299–303.
Zamboni P, Galeotti R, Menegatti E, et al. Chronic cerebrospinal
protocol. DS was responsible for participant selection. DKo undertookvenous insuﬃciency in patients with multiple sclerosis.
safety monitoring and recruitment. ALT, LM, KBK, DKL, DS, SL, PS,
RO, and DKo reviewed and analysed the data. YZ and IY did statistical J Neurol Neurosurg Psychiatry 2009; 80: 392–99.
Zamboni P, Galeotti R, Menegatti E, et al. A prospective open-label analysis, and ALT, LM, KBK, and DS did the literature searches. Study of endovascular treatment of chronic cerebrospinal venous insuﬃciency. J Vasc Surg 2009; 50: 1348–58.
Conﬂicts of interest W5’s ground-breaking investigation of CCSVI. http://www.ctvnews.