How are you taking part in this consultation?

You will not be able to change how you comment later.

You must be signed in to answer questions

    The content on this page is not current guidance and is only for the purposes of the consultation process.

    Outcome measures

    Efficacy and safety outcomes are included. Further details are provided below.

    Efficacy

    Identified outcomes relevant to efficacy include:

    • Headache response. Measures the degree of resolution or improvement of headaches after treatment. It can be assessed by:

    • Resolution rate indicating complete, partial or no resolution of headache.

    • Severity of headache. Assessed using tools such as the HIT-6, a questionnaire measuring the impact of headache on daily life. Scores range from 36 to 78, with higher scores indicating a greater impact of headache on a person's life. Impact levels are categorised as: little or no impact (49 or less), some impact (50 to 55), substantial impact (56 to 69), and several impact (60 to 68). The severity of headache is also measured using VAS, where 0 represents no pain and 10 represents the worst imaginable pain.

    • Number of headache days

    • Overall symptom resolution or improvement

    Captures the comprehensive spectrum of SIH-related symptoms, including headache, nausea, vomiting, and visual or cognitive disturbances, with either complete resolution or improvement. Patient-reported outcome scales such as PGIC can be used to assess overall symptom resolution or improvement. PGIC is a 7-point scale, ranging from "very much improved" to "very much worse".

    • Specific symptom improvement

    Improvement in SIH related symptoms improvement beyond headache such as tinnitus, dizziness or vertigo, cognitive dysfunction, and hearing disturbances.

    • Quality of life improvement. Measured using patient-reported outcome tools. These include:

    • VAS which scores from 0 to 10, where 0 represents very low quality of life and 10 represents very high quality of life.

    • SF-36 health survey which assess physical and mental health across eight domains, each scored from 0 to 100 with higher scores indicating better health status.

    • Imaging improvement

    Assessed using Bern score, a brain MRI based scoring system for SIH. It classifies people into high, intermediate, or low probability of identifying CSF leak. Score ranges from 0 to 9, with 0 to 2 indicating low probability, 3 to 4 indicating intermediate, and 5 to 9 indicating high probability. Imaging improvement is reflected by a reduction in the Bern score after treatment.

    Imaging improvement can also be assessed by MRI SIH score which evaluate brain MRI findings in people with SIH. The total score ranges from 0 to 9, with a higher score indicating more severe imaging abnormalities and 0 indicates no signs of SIH.

    Safety

    Identified outcomes relevant to safety include:

    • Rate of recurrence or retreatment

    This includes recurrence or residual fistulas at the original treatment site as well as the development of new fistulas at different spinal levels requiring further treatment. Persistent or recurrent symptoms such as tinnitus, hearing disturbance and dizziness may also be present.

    • Post-treatment rebound intracranial hypertension, or rebound headache

    This is a post-treatment effect presenting as new or different headaches within 24 to 48 hours after embolisation. These headaches are frequently associated with rebound intracranial hypertension and may require medication such as acetazolamide, and occasionally, invasive intervention, until resolution.

    • Perioperative technical complications

    These are intra and perioperative issues related to embolic material or catheter use. These include embolic agent migration, such as onyx migration into azygos vein or contamination of epidural or paraspinal veins. Other complications may include inadequate or unintended embolic agent deposition, including insufficient penetration, reflux, or extravasation into adjacent muscle, vessel injury and contrast leakage.

    • Localised pain at the embolisation site

    Population and studies description

    This interventional procedures overview is based on evidence from 1 systematic review and meta-analysis and 5 observational studies. The systematic review included a total of 321 people, 190 had the procedure. Observational studies included a total of 229 people, but there may be significant overlap between the study populations as three studies included in the evidence summary were also included in the systematic review. This is a rapid review of the literature ,and a flow chart of the complete selection process is shown in figure 1. This overview presents 6 studies as the key evidence in table 2 and table 3, and lists 9 other relevant studies in appendix B, table 5.

    Among the 5 observational studies, 2 were retrospective case series from the USA (Brinjikji, 2024; Parizadeh, 2023). The other 3 were prospective case series conducted in France (Cagnazzo, 2024; Cagnazzo, 2025; Goapper, 2025). All of these studies were single-centre studies, and none of them included UK centres. The 3 key studies (Brinjikji,2024; Parizadeh 2023; and Cagnazzo, 2024) included in the evidence summary were also included in the systematic review (Jazayeri, 2025). Additionally, there may be an overlap between the populations in 3 studies (Cagnazzo 2024; Cagnazzo, 2025; Goapper, 2025).

    Follow-up varied among studies. Most studies reported both imaging follow-up and clinical follow-up (Brinjikji, 2024; Parizadeh, 2023; Cagnazzo, 2024; Cagnazzo, 2025). Clinical follow-up ranged from 3 months (Parizadeh, 2023) to 15 months (Brinjikji, 2024), and imaging follow-up ranged from 3.3 months (Parizadeh, 2023) to 8 months (Cagnazzo, 2025).

    Study populations ranged from 18 people (Parizadeh, 2023) to 100 people (Brinjikji, 2023). All studies included people with SIH and confirmed CSFVFs. Female sex was the majority in all studies, ranging from 65% (Cagnazzo, 2025) to 72% (Parizadeh, 2023). Median or mean ages were consistently in the late 50s to early 60s, ranging from 59.2 years (Brinjikji, 2024) to 63 years (Cagnazzo, 2024). Four studies reported prior treatment before transvenous embolisation, either surgery or targeted or non-targeted EPB (Brinjikji, 2024; Parizadeh, 2023; Cagnazzo, 2024; Cagnazzo, 2025). Three studies reported presence of multiple fistulas (Brinjikji, 2023; Cagnazzo, 2024; Cagnazzo 2025). Table 2 presents study details.

    Figure 1 Flow chart of study selection

    Table 2 Study details

    Study no.

    First author, date

    country

    Characteristics of people in the study (as reported by the study)

    Study design

    Inclusion criteria

    Intervention

    Follow up

    1

    Brinjikji

    2024

    USA

    N=100

    Mean age=59.2 (SD 10.9)

    Female sex=68 (68%)

    Number of fistulas=105

    Prior treatment:

    Surgical treatment=9

    Epidural blood patch=59

    Retrospective case series

    (June 2020 to June 2022)

    People with SIH confirmed to have CSFVF evaluated by brain MRI with and without gadolinium contrast, spine MRI, and advanced myelography including either DSM or CT myelography

    Transvenous embolisation

    Paraspinal vein catheterisation and embolisation of fistula using Onyx (Onyx 34 and Onyx 18)

    Mean clinical follow-up 15 months (range 6.8 months)

    Mean imaging follow-up 8.3 months (range 7.7 months)

    2

    Parizadeh 2023

    USA

    N=18

    Median age 60 (IQR range 50 to 65)

    Female sex=13 (72%)

    Number of fistulas=18

    Prior treatment:

    Targeted or non-targeted epidural blood/fibrin patches=11

    Surgery for Chiari I malformation=3

    Retrospective case series

    People with SIH having CSFVF. SIH diagnosed based on clinical and/or brain MRI findings

    Transvenous embolisation

    Transvenous catheterisation and embolisation at foraminal venous plexus using Onyx (18 and 34), often with balloon and/or coil assisted "pressure cooker" method for improved penetration and reduce reflux

    Clinical follow-up 3 months (range 2.3 to 5.0)

    Imaging follow-up 3.3 months (range 2.3 to 4.7)

    3

    Cagnazzo 2024

    France

    N=21

    Median age=63 (IQR range 58 to 71)

    Female sex=15 (71.5%)

    Number of fistulas=30

    Prior treatment:

    Epidural blood patching=7

    Prospective case series

    November 2022 to January 2024

    People with SIH having confirmed CSFVF through DSM. SIH diagnosed based on clinical symptoms and MRI

    Transvenous embolisation

    Transvenous catheterisation and embolisation of CSFVF using Onyx 18, with coverage of target and adjacent epidural plexus levels.

    Mean clinical follow-up 9 months (SD 6.4)

    Mean imaging follow-up 7 months (SD 5.1)

    4

    Cagnazzo, 2025

    France

    N=60

    Mean age=61 years (SD 13)

    Female sex=26 (65%)

    Prior treatment:

    Non-targeted EBP=13

    Prospective case series

    November 2022 to December 2024

    People with SIH having confirmed through DSM. SIH diagnosed based on clinical symptoms and brain MRI signs based on SIH score

    Transvenous embolisation

    Transvenous catheterisation and embolisation of CSFVF using Onyx 18, with coil occlusion of venous connection to prevent Onyx migration

    Mean clinical follow-up 12 months (SD 4.4)

    Mean imaging follow-up 8 months (SD 4.1)

    5

    Goapper 2025

    France

    N= 30

    Mean age= 60.4 years (SD 14.1)

    Female sex=20 (66.6%)

    Number of fistulas=10 people with multiple fistulas

    Prospective case series

    People with SIH and a definitive diagnosis of CSFVF through DSM. SIH diagnosed based on clinical symptoms and brain MRI

    Transvenous embolisation

    Transvenous catheterisation and embolisation of CSFVF using Onyx 18, with coil to blog high-risk venous connection.

    Mean clinical and imaging follow-up 10 months (range 6 to 24)

    QoL follow-up 3 months

    6

    Jazayeri 2025

    Multiple countries

    6 studies relevant to CSFVF, n=190 people

    There is likely overlap with other studies included in this overview.

    The following studies reported in this systematic review have also been included separately in this overview:

    Brinjikji (2024)

    Parizadeh (2023)

    Cagnazzo (2024)

    Systematic review and Meta-analysis

    Adults with confirmed CSFVSs on imaging treated either surgically or through transvenous embolisation

    Transvenous embolisation

    3 to 15 months

    Table 3 Study outcomes

    First author, date

    Efficacy outcomes

    Safety outcomes

    Brinjikji

    2024

    Headache response

    • Resolution/improvement of headache:

    Resolution of headache=75% (72/96)

    Improvement of headache=19.8% (19/96)

    No change/ worsening=5.2% (5/96)

    Overall symptom resolution/improvement

    • Complete symptom resolution=58% (58/100)

    • Significant improvement=37% (37/100)

    • No improvement=5% (5/100)

    • Sub-group analysis

    Resolution by sex:

    Male=43.1%; improvement=16.2%; no improvement=20%, p=0.02

    No history of headache before leak onset:

    Resolution=82.8%; improvement=62.2%; no improvement=40%, p=0.0004

    Mean duration of symptoms before treatment:

    Resolution=32 months (SD 39.4); improvement=49.2 (SD 16.6); no improvement=82.6 (SD 81.2)

    Suspected RIH requiring acetazolamide at 3 months:

    Resolution=5.2%; improvement=29.7%; no improvement=60%, p=0.001

    Specific symptom improvement

    • Tinnitus

    Resolution=64% (35/55)

    Improvement=29% (16/55)

    • Cognitive dysfunction

    Resolution=66% (29/44)

    Improvement=27% (12/44)

    • Impaired hearing

    Resolution=71% (25/35)

    Improvement=23% (8/35)

    • Dizziness/vertigo

    Resolution=76% (22/29)

    Improvement=17% (5/29)

    • Visual alteration

    Resolution=60% (9/15)

    Improvement=40% (6/15)

    • Gait disturbance

    Resolution=83% (10/12)

    Improvement=0

    Imaging improvement

    • Bern SIH score:

    Mean post-treatment score=0.9 (SD 1.6)

    Improved Bern score=81%

    No change=3%

    Worsening = 0

    Bern SIH score of 0 at baseline=14%

    Rate of recurrence/retreatment

    • Retreatment: 17/100;

    Development of new CSFVF at new level: 10 people

    Recurrent/residual leak at site of embolisation: 7 people

    Post-treatment rebound intracranial hypertension/rebound headache

    • Total with suspected RIH: 19/100 (all required acetazolamide, 2 further needed invasive treatment

    Perioperative technical complications

    • Paraspinal or epidural vein perforation: 5/100

    • Severe postoperative pain requiring narcotic, gabapentin or steroid injection: 3/100

    • Radiculopathy/numbness: 1/100

    Parizadeh 2023

    Headache response

    • Severity of headache

    HIT-6 score change:

    Before treatment=68 (range 62 to 72)

    After treatment=36 (range 36 to 38), p <0.001

    HIT-6 improvement in 94% and significantly improved in 81% post-embolisation.

    Overall symptom resolution/improvement

    • Patient Global Impression of Change (PGIC) Scale

    Any improvement=94% (17/18)

    Very much/much improvement=89% (16/18)

    Imaging improvement

    • Bern score change:

    Before treatment=8 (range 6 to 8)

    At follow-up= 3 (range 1.5 to 3.5), p< 0,001

    Rate of recurrence/retreatment: 2/18

    Post-treatment rebound intracranial hypertension/rebound headache

    • Post-embolisation headache (suggestive of RIH): 13/18; 9 treated with acetazolamide

    Perioperative technical complications

    • Complication rate: 22% (4/18)

    • Intramuscular extravasation of Onyx: 2/18

    • Transient epidural contrast extravasation identified on microcatheter injection: 1/18

    • Mild Onyx reflux into the azygos vein: 1/18

    Localised pain at the site of embolisation: 15/18

    Cagnazzo 2024

    Headache response

    • Severity of headache

    Mean HIT-6 score:

    Before treatment=67 (SD 15)

    After treatment=38 (SD 9), p< 0.0001

    Mean VAS score:

    Before treatment=8 (SD 1.9)

    After treatment=1.2 (SD 2), p<0.0001

    • Number of headache days in a month

    Before treatment=23.5 (SD 10)

    After treatment=3.2 (SD 6.6), p<0.0001

    Overall symptom resolution/improvement

    • Complete clinical recovery=67% (14/21)

    • Substantial/partial improvement=28% (6/21)

    Specific symptom improvement

    • Dizziness/vertigo=69% (improved) (11/16)

    • Fatigue=82% (resolved) (14/17)

    • Tinnitus=60% (resolved) (9/15)

    Quality of life

    • Mean VAS score for QoL

    Before treatment=2.6 (SD 2.5)

    After treatment=8.6 (SD 1.8), p<0.0001

    Imaging improvement

    • Mean SIH score:

    Before treatment=6 (SD 2.5)

    After treatment=1.4 (SD 1.6), p< 0.0001

    Rate of recurrence/retreatment: 2/21

    Post-treatment rebound intracranial hypertension/rebound headache

    • Post-treatment RIH: 7/21; 6 treated with acetazolamide

    Perioperative technical complications

    • Complication rate: 28% (6/21)

    • Inadequate penetration of embolic material: 2

    • Asymptomatic migration of embolic material into the azygos vein: 2

    Localised pain at the site of embolisation: 4

    Cagnazzo, 2025

    Headache response

    Resolution/improvement at follow-up (3 months)

    • Orthostatic headache=89% (25/28)

    • Non-orthostatic headache=50% (2/4)

    • Activity related headache=83% (24/29)

    • Valsalva related headache=84% (21/25)

    Overall symptom resolution/improvement

    • Follow up (3 months):

    Significant clinical recovery=95% (38/40);

    Complete clinical recovery=86.8% (33/38)

    Specific symptom improvement Resolution/improvement at follow-up (3 months)

    • Fatigue=91% (32/35)

    • Tinnitus=61% (11/18)

    • Dizziness=77% (20/26)

    Imaging improvement

    • Median SIH score:

    Before treatment=6 (IQR range 4 to 8)

    After treatment (24 hours)=3.5 (IQR range 2 to 5) (p=0.01)

    3 months follow-up=2 (IQR range 0 to 4)

    (p=0.004)

    Rate of recurrence/retreatment

    • Retreatment: 4/40

    New CSFVF: 3 patients

    Incomplete occlusion from initial procedure: 1 patient

    Post-treatment rebound intracranial hypertension/rebound headache

    • Post-treatment rebound headache: 32.5% (13/40), 9 treated with acetazolamide

    Perioperative technical complications

    • Asymptomatic migration of embolic material into the azygos vein: 2/40

    • Microcatheter perforation of epidural plexus: 2/40

    Localised pain at the site of embolisation: 9/40

    Goapper, 2025

    Headache response

    • Resolution of headache

    Complete headache resolution=64% (18/30)

    Partial headache resolution=35% (10/30)

    • Severity of headache

    VAS score (0 to 100):

    Before treatment (mean)=74.3 (SD 27.6)

    At follow-up (mean)=9.0 (SD 15.6)

    Significant reduction (p<0.0001)

    • Number of headache days per month

    Before treatment (mean)=25 (SD 9.2)

    At follow-up (mean)=2.8 (SD 7.4)

    Significant reduction (p<0.0001)

    Specific symptom improvement

    • Dizziness/Vertigo:

    Complete resolution=42% (8/19)

    Partial resolution=37% (7/19)

    • Tinnitus:

    Complete resolution=38% (5/13)

    Partial resolution=23% (3/13)

    • Hearing disturbance:

    Complete resolution=25% (3/12),

    Partial resolution=33% (4/12)

    Quality of life

    • Physical health (SF-36 physical domain):

    Before treatment=82.3% rated their physical QoL as poor

    At follow-up (3 months)=70.6% rated their physical QoL as excellent

    • Mental health (SF-36 mental domain):

    Before treatment=76.4% rated their mental health as poor

    At follow-up (3 months)=100% rated their mental health as good to excellent

    • Overall quality of life (VAS-QoL):

    Before treatment = Mean score: 23.5 (SD 25.8)

    After treatment = Mean score: 88.7 (SD 15.5)

    High statistical significance (p<0.001)

    Rate of recurrence/retreatment

    • Recurrent symptoms: 3/30

    New CSFVF requiring second embolisation

    Post-treatment rebound intracranial hypertension/rebound headache

    • Post-treatment rebound headache: 40% (12/30)

    Perioperative technical complications

    • Asymptomatic migration of embolic material into the azygos vein: 1/30

    • Epidural plexus perforation: 2/30

    Localised pain at the site of embolisation: 7/30

    Jazayeri SB, 2025

    Headache response

    • Resolution of headache:

    Complete or partial resolution of headache

    93.9% (95% CI: 88.3%–96.9%);

    Complete headache resolution rate

    73.3% (95% CI: 65.1% to 80.2%)

    Partial headache response

    20.6% (95% CI14.5% to 28.4%)

    • Severity of headache:

    HIT-6 score mean change -29.1 points (95% Cl -32.7 to -25.5)

    Overall symptom resolution

    • Complete or partial resolution

    93.9% (Cl 88.3% to 96.9%);

    Complete resolution of all symptoms

    59.1% (95% Cl 50.5% to 67.1%)

    Partial response to all symptoms

    34.8% (95% Cl 27.2% to 43.3%)

    Imaging improvement

    • Bern score mean change

    - 4.7 points (95% CI -5.3 points to -4.2 points) (p<0.01)

    Rate of recurrence/retreatment: 15.3% (95% CI 10.3% to 22.1%)

    Post-treatment rebound intracranial hypertension/rebound headache

    • Suspected RIH: 48.9% (95% CI 21.5% to 76.9%)

    Procedure technique

    Of the 5 observational studies, all provided at least some detail of the transvenous embolisation technique, with four giving comprehensive detail, including devices used (Brinjikji, 2024; Parizadeh, 2023; Cagnazzo, 2024; Cagnazzo, 2025). Across all studies, the procedure was performed under general anaesthesia. The venous access was obtained through the common femoral vein (Brinjikji,2024; Cagnazzo,2024; Cagnazzo, 2025; Goapper, 2025) or upper extremity veins (Parizadeh, 2023). Microcatheter, such as the Headway Duo or Scepter XC was used for microcatheterisation. Embolisation was predominantly performed with Onyx 2018 across all studies. In 2 studies (Brinjikji, 2024; Parizadeh,2023), Onyx 34 was first used to create a plug near the vein and then switched to Onyx 18 to continue embolisation. Three studies reported using adjunctive techniques such as balloon microcatheters and coils to reduce reflux in some cases (Cagnazzo, 2025; Goapper, 2025) and optimise embolic penetration, referred to as the "pressure cooker" technique (Parizadeh, 2023). This technique creates a proximal plug to prevent reflux, enabling controlled forward injection of embolic agent into the targeted fistula.

    Efficacy

    Headache response

    Headache response after transvenous embolisation was reported across all studies.

    Resolution or improvement in headache was reported in 4 studies. The pooled analysis of 4 studies (N=131) in Jazayeri (2025) reported complete or partial resolution of headache in 93.9% (95% Cl 88.3% to 96.9%), with complete resolution rate in 73.3% (95% Cl 65.1% to 80.2%) and partial headache resolution in 20.6% (95% Cl 14.5% to 28.4%). Three observational studies also reported similar headache resolution results.

    Brinjikji (2024) observed complete resolution in 75% (72/96), partial improvement in 19.8% (19/96), and no change in 5.2% (5/96) over a mean follow-up period of 15 months (SD 6.8). Goapper (2025) reported complete resolution in 64% (18/30) and partial resolution in 35% (10/30) over a mean follow-up of 10 months (range 6 to 24). Cagnazzo (2025) found orthostatic headache resolved or improved in 89% (25/28), activity-related headache in 83% (24/29), Valsalva-related headache in 84% (21/25), and non-orthostatic headache in 50% (2/4) at 3-month follow-up.

    Headache severity improvement was reported in 4 studies. The pooled mean HIT-6 score change from three studies (N=76) in Jazayeri (2025) was –29.1 (95% Cl –32.7 to –25.5). Parizadeh (2023) reported a statistically significant reduction (p less than 0.001) in HIT-6 score from 68 (range 62 to 72) to 36 (range 36 to 38). HIT-6 score was improved in 94% of people and significantly improved in 81% post-procedure at a median follow-up of 3 months (range 2.3 to 5). Cagnazzo (2024) also observed a statistically significant decrease (p less than 0.0001) in HIT-6 score from 67 (SD 15) to 38 (SD 9). Similarly, mean VAS score significantly decreased (p less than 0.0001) from 8 (SD 1.9) to 1.2 (SD 2) over a mean follow-up period of 9 months (SD 6.4). Goapper reported a significant decrease (p less than 0.0001) in VAS (0-100) score from 74.3 (SD 27.6) to 9.0 (SD 15.6).

    Reduction in headache per month was reported in two studies. Cagnazzo (2024) reported a statistically significant decrease (p less than 0.0001) in headache days per month from 23.5 (SD 10) to 3.2 (SD 6.6). This was consistent with Goapper (2025), who reported a statistically significant decrease (p less than 0.0001) in headache days per month from 25 (SD 9.2) to 2.8 (SD 7.4).

    Overall symptom resolution/improvement

    Overall symptom resolution or improvement of SIH following transvenous embolisation was reported in 5 studies. The pooled analysis of 4 studies (N =132) in Jazayeri (2025) reported a combined complete or partial symptoms resolution rate of 93.9% (Cl 88.3% to 96.9%), with complete resolution in 59.1% (95% Cl 50.5% to 67.1%) and partial improvement in 34.8% (95% Cl 27.2% to 43.4%).

    Brinjikji (2024) reported complete symptom resolution in 58% (58/100) of patients, improvement in 37% (37/100) and no improvement in 5 people over a mean clinical follow-up of 15 months (SD 6.8).

    Brinjikji (2024) performed subgroup analysis to identify predictors of better outcomes (symptom resolution). People with complete resolution were more often male (43.1%), more likely to have no history of headache before the onset of the leak (82.8%). Mean symptom duration before treatment was shorter in the resolution group, 32 months (SD 39.4), compared with the improvement group, 49.2 months (SD 61.6) and the no improvement group, 82.6 months (SD 81.2). RIH requiring medication occurred in 5.2% of resolution patients, 29.7% in the improvement group and 60% in the no improvement group.

    Parizadeh (2023) assessed overall symptom resolution using the PGIC scale. The study reported 94% (17/18) of people experienced some improvement in SIH symptoms, with "much or very much improvement" in 89% (16/18) at 3 months follow-up.

    In Cagnazzo (2024), complete symptom resolution was achieved in 67% (14/21) and partial improvement in 28% (6/21) at a mean follow-up of 9 months (SD 6.4). One patient with no overall improvement had a chronic headache with a history of SIH for more than 15 years.

    Cagnazzo (2025) reported overall symptom improvement in 95% of people (38/40), with complete resolution in 86.8% (33/38) at 3 months follow-up. The two people without any improvement had a longstanding non-orthostatic headache and tinnitus for over 3 years.

    Specific symptom resolution/ improvement

    Specific symptom resolution/improvement after transvenous embolisation was reported in 4 studies.

    Tinnitus was the most frequently reported symptom. Resolution rates were 64% (35/55) in Brinjikji (2024), 60% (9/15) in Cagnazzo (2024), 61% (11/18) in Cagnazzo (2025) and 38% (3/13) in Goapper (2025). Improvement without complete resolution occurred in 29% (16/55) in Brinjikji (2024), 23% (3/13) in Goapper (2025).

    Dizziness or vertigo was reported to be completely resolved in 76% (22/29) in Brinjikji (2024), 69% (11/16) in Cagnazzo (2024), 77% (20/26) in Cagnazzo (2025), and 42% (8/19) in Goapper (2025). Partial resolution was reported in 17% (5/29) and 37% (7/19) in Goapper (2025).

    Fatigue was assessed in 2 studies, with resolution rates of 82% (14/17) in Cagnazzo (2024) and 91% (32/35) in Cagnazzo (2025).

    Hearing disturbance resolved in 71% (25/35) in Brinjikji (2024) and 25% (3/12) in Goapper (2025), with partial improvement in 23% (8/35) in Brinjikji (2024) and 33% (4/12) in Goapper (2025).

    Brinjikji (2024) also reported other symptom resolution/improvement. Cognitive dysfunction was resolved in 66% (29/44) and improved in 27% (12/44). Visual alteration was resolved in 60% (9/15) and improved in 40% (6/15), and gait disturbance was resolved in 83% (10/12).

    Quality of life

    Quality of life outcomes after transvenous embolisation were reported in 2 studies.

    Cagnazzo (2024) assessed QoL using the VAS score (scale 0 to 10). There was significant improvement (p less than 0.0001) from a mean baseline VAS score of 2.6 (SD 2.5) to 8.6 (SD 1.8) post-procedure.

    Goapper (2025) assessed both physical and mental health domains using the SF-36 at baseline and at 3 months follow-up. In the physical health domain, 82.3% rated their physical QoL as poor at baseline compared with 70.6% rating it as excellent after treatment. In the mental health domain, 76.4% reported poor mental health at baseline, whereas 100% rated their mental health as good or excellent at follow-up. Additionally, an overall QoL was measured by the VAS scale (0 to 100). The mean VAS score significantly increased (p less than 0.001) from 23.5 (SD 25.8) to 88.7 (15.5) after treatment.

    Imaging improvement

    Imaging improvement after transvenous embolisation was reported in 5 studies. The pooled data from 4 studies (N=145) in Jazayeri (2025) reported a significant reduction (p less than 0.01) in post-treatment Bern score. (mean change -4.7, 95% Cl –5.3 to –4.2).

    Out of 4 observational studies reporting imaging improvement, 3 studies evaluated changes in the Bern SIH score. Brinjikji (2024) found a reduction of 5.9 (SD 3.3) to 0.9 (SD 1.6) at a mean imaging follow-up of 8.3 months (SD 7.7). 14 out of 100 people had a score of 0 at baseline, 81 showed improvement and 3 had no change in their Bern SIH score.

    Parizadeh (2023) reported a statistically significant (p less than 0.001) decrease from a median of 8 (range 6 to 8) to 3 (range 1.5 to 3.5) at a median follow-up of 3.3 months (range 2.3 to 4.7). Post-treatment, 83% (15/18) of people had improvement in SIH findings, while the score did not change in 3 people among whom one already had a low Bern SIH score.

    Cagnazzo (2024) observed a statistically significant (p less than 0.0001) reduction from 6 (SD 2.5) to 1.4 (SD 1.6) over a mean imaging follow-up period of 7 months (SD 5.1).

    One study used the SIH score. Cagnazzo (2025) reported a statistically significant (p = 0.004) reduction from a median of 6 (IQR 4 to 8) to 2 (IQR 0 to 4) at 3 months follow-up.

    Safety

    Rate of recurrence/retreatment

    The rate of recurrence/ retreatment was reported across all studies. Four studies in Jazayeri (2025) reported the rate of recurrence/retreatment. An overall rate of recurrence/retreatment after transvenous embolisation was 15.3% (95% Cl 10.3% to 22.1%).

    Brinjikji (2024) reported a retreatment rate of 17% (17/100) during a mean follow-up of 15 months (SD 6.8). Among these 17 patients, 10 developed new CSFVFs at different spinal levels, while 7 had recurrent or residual leaks at the site of embolisation.

    Parizadeh (2023) reported a treatment failure rate of 11% (2/18) during 3 months of follow-up.

    Cagnazzo (2024) reported a recurrence/retreatment rate of 9.5% (2/21) over a mean clinical follow-up period of 9 months (SD 6.4). Both required additional treatment; one for recurrent symptoms due to a small fistula, and one for incomplete symptom resolution after the initial procedure.

    Cagnazzo (2025) reported the recurrence or retreatment rate of 10% (4/40) during a mean clinical follow-up of 12 months (SD 4.4). All 4 underwent a second embolisation. 3 developed new CSFVFs, and one required an additional Onyx injection for incomplete occlusion from the initial procedure.

    Goapper (2025) reported a recurrence rate of 10% (3/30) during a mean follow-up of 10 months. All 3 required a second embolisation due to new CSFVFs.

    Post-treatment rebound intracranial hypertension/ rebound headache

    Post-treatment RIH/ rebound headache was reported across all studies. Four studies (N=131) in Jazayeri (2025) reported a rate of suspected RIH of 48.9% (95% Cl 21.5% to 76.9%).

    Brinjikji (2024) reported a suspected RIH rate of 19% (19/100). All 19 people had acetazolamide for a median duration of 3 weeks, and 2 required additional invasive treatment.

    Parizadeh (2023) reported that 72% of people (13/18) developed post-treatment headache suggestive of RIH. 9 people were treated with acetazolamide, and symptoms were resolved within 3 months in all except 2 people.

    Cagnazzo (2024) reported a post-treatment RIH rate of 33% (7/21). Headaches typically appeared within 24 to 48 hours after treatment and lasted a mean of 7.4 days (SD 3.7). Among 7 people, 6 were treated with acetazolamide for about 7 days until symptom regression. The other person was at high risk due to a large T12-L1 fistula and longstanding symptoms and received acetazolamide 3 days before and 14 days after the procedure.

    Cagnazzo (2025) reported a post-treatment rebound headache rate of 32.5% (13/40). Headache appeared within 24 to 48 hours after treatment and lasted a mean of 7 days (SD 3.5). Among 13 people, 9 were treated with acetazolamide for approximately a week until symptom resolution.

    Goapper (2025) reported a post-treatment rebound headache rate of 40% (12/30).

    Perioperative technical complications

    Technical complications during or after transvenous embolisation were reported across all 5 observational studies.

    The most frequent complication was asymptomatic Onyx migration into the azygos vein: Parizadeh (2023) in 1/18 patients, Cagnazzo (2024) in 2/21 patients, Cagnazzo (2025) in 2/40 patients, and Goapper (2025) in 1/30.

    Microcatheter or epidural plexus perforation was reported in Cagnazzo (2025) in 2/40 and Goapper (2025) in 2/30 people, without any long-term complications.

    Extravasation or leakage of Onyx or contrast was reported in Parizadeh (2023) with intramuscular Onyx extravasation in 2 people and transient epidural contrast leakage in 1 patient. Cagnazzo (2024) reported inadequate Onyx penetration in 2 patients.

    Other complications were reported in Brinjikji (2024). In 100 patients, paraspinal or epidural vein perforation was observed in 5 people, severe postoperative pain requiring medication in 3 people and postoperative radiculopathy or numbness in 1 patient.

    Localised pain at the site of embolisation

    Localised pain at the site of embolisation was reported in 4 studies. Parizadeh (2023) observed a high rate of localised pain at 83% (15/18). In cases where the adjunctive "pressure cooker" technique was used, transient localised back pain was significantly more frequent compared with conventional embolisation (100% versus 50%, p=0.02).

    Cagnazzo (2024) reported a localised pain rate of 19% (4/21), which resolved within 2 weeks. In Cagnazzo (2025), 22.5% of people (9/40) experienced localised pain, with a mean duration of 3 days. Goapper (2025) reported a rate of 23% (7/30).

    Anecdotal and theoretical adverse events

    Expert advice was sought from consultants who have been nominated or ratified by their professional society or royal college. They were asked if they knew of any other adverse events for this procedure that they had heard about (anecdotal), which were not reported in the literature. They were also asked if they thought there were other adverse events that might possibly occur, even if they had never happened (theoretical).

    They listed the following anecdotal adverse events:

    • Localised pain to the site of injection

    • Rebound intracranial hypertension

    • Rebound headache

    • CSFVF recurrence

    • Need for retreatment

    • Venous perforation

    • Pulmonary emboli

    • Unintended extensive embolic agent embolisation of the venous plexus near fistula.

    • General risks of endovascular procedures, including bleeding, infection, risk of general anaesthesia and vessel injury.

    7 professional expert questionnaires for this procedure were submitted. Find full details of what the professional experts said about the procedure in the https://www.nice.org.uk/guidance/ipg10318/documents.

    Validity and generalisability

    • Sample sizes were small across all studies, reflecting the recent recognition of CSFVF in SIH and the limited use of this procedure in clinical practice. The largest sample size was 100 in Brinjikji (2024), while the smallest sample size was 18 in Parizadeh (2023).

    • None of the studies were conducted in the UK. Study centres were based in the USA (Brinjikji, 2024; Parizadeh, 2023) and France (Cagnazzo, 2024; Cagnazzo, 2025; Goapper, 2025). All 3 French studies originated from the same institute, Montpellier University Hospital, raising the possibility of patient overlap between the studies.

    • All the included studies had early to mid-term follow-up ranging from 3 months to 15 months. Parizadeh (2023) only included people with more than 3 months of clinical and imaging follow-up. Similarly, Cagnazzo (2024) and Cagnazzo (2025) mandated complete baseline and 3-month follow-up MRI and clinical data. There was a lack of long-term follow-up.

    • Across all studies, people selection was based on a confirmed diagnosis of SIH and definite CSFVF identified on advanced imaging (DSM or CT myelography) before undergoing transvenous embolisation. Cagnazzo (2024) and Cagnazzo (2025) required multidisciplinary team approval before treatment, which was not explicitly mentioned in other studies.

    • All studies were observational studies, either retrospective or prospective case series and did not involve random assignment. None of the studies reported adjustment to potential confounders (symptom duration, baseline severity, comorbidities), making them susceptible to bias. Additionally, there was a difference in measurement tools for assessing clinical and imaging outcomes. For example, imaging improvement was assessed using the Bern SIH score in 2 studies (Brinjikji, 2024; Parizadeh, 2023) and the MRI SIH score in 2 studies (Cagnazzo, 2024; Cagnazzo, 2024). There was no standardised outcome definition across studies.

    • The overall procedure of transvenous embolisation, including transvenous access, microcatheter navigation, and use of embolic agent (primarily Onyx 18), was consistent across studies. Parizadeh (2023) used a "pressure cooker" technique with balloon catheters and coils to enhance embolic penetration. In some cases, coils were also used to prevent Onyx migration (Cagnazzo 2024; Cagnazzo 2025; Goapper 2025).

    • Across the included studies, no external funding sources or manufacturer involvement was reported, except in Brinjikji (2024). In that study, one author disclosed multiple industry affiliations, including with Medtronic, and financial interests, while all other authors declared no conflicts. The same author was also involved in Jazayeri (2025).

    • Parizadeh (2023) reported a high rate of localised pain at the site of embolisation 83% compared with other studies, which ranged from 19% to 23% (Cagnazzo, 2024; Cagnazzo, 2025; Goapper, 2025). Additionally, Parizadeh (2023) observed local pain in all cases who underwent treatment using the "pressure cooker" technique. The study identified transient localised back pain as a side effect of this technique.

    • Parizadeh (2023) reported a high rate of RIH post-treatment, occurring in 72% (13/18), which was higher than rates reported in other studies, where RIH ranged from 19% to 40% (Brinjikji,2024; Cagnazzo,2024; Cagnazzo, 2025; Goapper, 2025). This increased incidence was associated with the "pressure cooker" technique, which resulted in RIH in 67% compared to17% with the conventional technique. The study suggested that the higher rate of RIH may be due to more complete CSFVF occlusion, leading to stronger rebound headaches. Another contributing factor could be a lowered threshold for treating RIH with acetazolamide over time, resulting in a higher number of recorded cases.

    • All included studies reported that transvenous embolisation was a safe and effective treatment for CSFVF. The one systematic review (Jazayeri, 2025), which compared transvenous embolisation with surgery reported the procedure to have comparable safety and efficacy outcomes when compared with surgery. Technical refinement, such as the pressure cooker technique, might potentially improve the completeness of embolisation. Early diagnosis and treatment were found to be associated with a better likelihood of complete recovery (Brinjikji, 2024; Goapper, 2025). There is a lack of RCTs, and current evidence is based on small-sized prospective and retrospective case series with early to mid-term follow-up.

    Ongoing trials

    • A Clinical Registry of Spontaneous Intracranial Hypotension.

    NCT05922514 . N= 200. USA. Expected completion: July 2026

    • Transvenous Approach for the Treatment of Cerebral Arteriovenous Malformations. NCT03691870 . N = 77. Canda. Completion: July 2025. (Completed)