Original Research
Neuroradiology/Head and Neck Imaging
October 12, 2017

CSF Venous Fistulas in Spontaneous Intracranial Hypotension: Imaging Characteristics on Dynamic and CT Myelography

Abstract

OBJECTIVE. The objective of this study is to describe the anatomic and imaging features of CSF venous fistulas, which are a recently reported cause of spontaneous intracranial hypotension (SIH).
MATERIALS AND METHODS. We retrospectively reviewed the records of patients with SIH caused by CSF venous fistulas who received treatment at our institution. The anatomic details of each fistula were recorded. Attenuation of the veins involved by the fistula was compared with that of adjacent control veins on CT myelography (CTM). Visibility of the CSF venous fistula on CTM and a modified conventional myelography technique we refer to as dynamic myelography was also compared.
RESULTS. Twenty-two cases of CSF venous fistula were identified. The fistulas were located between T4 and L1. Ninety percent occurred without a concurrent epidural CSF leak. In most cases (82%), the CSF venous fistula originated from a nerve root sleeve diverticulum. On CTM, the abnormal veins associated with the CSF venous fistula were seen in a paravertebral location in 45% of cases, centrally within the epidural venous plexus in 32%, and lateral to the spine in 23%. Differences in attenuation between the fistula veins and the control veins was highly statistically significant (p < 0.0001), with a threshold of 70 HU perfectly discriminating fistulas from normal veins in our series. When both CTM and dynamic myelography were performed, the fistula was identified on both modalities in 88% of cases.
CONCLUSION. CSF venous fistulas are an important cause of SIH that can be detected on both CTM and dynamic myelograph y and may occur without an epidural CSF leak. Familiarity with the imaging characteristics of these lesions is critical to providing appropriate treatment to patients with SIH.
CSF venous fistulas are a recently discovered cause of spontaneous intracranial hypotension (SIH) [1, 2]. In this entity, there is a direct connection between the spinal subarachnoid space and a paraspinal vein that permits unregulated loss of CSF. Detection of CSF venous fistulas is important because they represent a treatable cause of SIH in patients who have no other visible CSF leak on spinal imaging [13].
Only a few previously reported cases of CSF venous fistulas exist, and most of these were identified in patients receiving general anesthesia and undergoing digital subtraction myelography (DSM), a modification of conventional myelography technique that is not performed at all institutions [2]. Determining how to localize these CSF venous fistulas on CT myelography (CTM) would be desirable because CTM is considered by many authors to be the test of choice in evaluating CSF leaks and is the first spinal imaging test performed for patients with SIH at many institutions [4]. One small series of patients who underwent imaging with CTM described an imaging sign known as hyperdense paraspinal vein sign that indicates the presence of a CSF venous fistula [5]. However, experience with this imaging sign, and with CSF venous fistulas in general, currently is quite limited.
The purpose of this study is to describe the anatomic location and imaging appearance of CSF venous fistulas diagnosed at our institution. We sought to compare the appearance of CSF venous fistulas on CTM with its appearance on a variant of conventional myelography that does not require general anesthesia or digital subtraction capability, which we describe herein and refer to as dynamic myelography.

Materials and Methods

This investigation is a retrospective cross-sectional study of patients who had CSF venous fistulas diagnosed. The investigation was approved by the institutional review board at Duke University Medical Center and is compliant with HIPAA regulations.

Subjects

Subjects were identified from a review of patients referred for possible SIH between September 2013 (the date of our first identified CSF venous fistula) and January 2017. A CSF venous fistula was determined to be present if imaging showed clear contrast filling of a paraspinal vein on conventional myelography or a positive hyperdense paraspinal vein sign on CTM [5].
Demographic information from the electronic medical record was recorded. Each patient completed a Headache Impact Test (HIT-6) questionnaire before treatment. The HIT-6 is a six-question validated headache severity instrument with a score range of 36–78 points [6, 7]. Scores greater than 60 indicate a very severe effect on life; scores from 56 to 59, a substantial impact; scores of 50–55, some impact; and scores of 49 or less, little or no impact. The CSF opening pressure measured during the patient's first myelographic procedure was also obtained from the medical record. Details of treatment with epidural patching, surgery, or both were recorded.

Procedure Technique

At our institution, CTM is the first-line test for evaluation of a suspected CSF leak. It is performed immediately after CT fluoroscopy–guided lumbar puncture, with use of the technique previously described elsewhere [8]. In brief, contrast material is injected while the patient is on the CT table, and the patient is assisted in performing a pelvic lift to distribute the contrast material cranially and then is scanned immediately, typically within 2–3 minutes after contrast injection. The purpose of rapid image acquisition is to image high-flow leaks before contrast medium has a chance to spread too far from the leak site in the epidural space and to capture subtle low-flow leaks that may become less apparent on delayed imaging. Imaging is performed using an MDCT scanner (Discovery CT750 HD, GE Healthcare) with use of the following scan parameters: detector configuration, 32 × 0.625; helical scan mode; rotation time, 0.8 second; pitch, 0.969; tube voltage, 120 kVp; automatic exposure control on; tube current, 300–800 mA; noise index, 19.5; image thickness, 2.5 mm; interval, 2.5 mm; and reconstruction thickness, 0.625 mm.
If the initial imaging examination (performed either at our institution or elsewhere) suggests a high-flow CSF leak or a CSF venous fistula, we perform myelography under fluoroscopy with use of a technique we refer to as dynamic myelography. In this technique, the patient is positioned on a tilting C-arm fluoroscopy table so that the suspected leak site lies along the dependent side of the thecal sac (i.e., patients with suspected CSF venous fistulas or leaks arising from the nerve root are placed in a lateral decubitus position, whereas patients with suspected ventral leaks are placed in a prone position). A lumbar puncture is performed, and once the subarachnoid space is accessed, the patient is tilted into the feet-down position. A total of 10 mL of contrast material (Isovue-M 300, Bracco) is injected slowly to ensure that it remains contained in the lumbosacral region. The needle is then removed, and the patient is maintained in the prone or decubitus position while the table is slowly tilted from the feet-down position into the head-down position, to maintain a tight bolus of contrast material. Migration of contrast material is observed using intermittent fluoroscopy, with spot images obtained once the contrast material migrates to the level of the suspected leak. Table tilt is adjusted to maintain the contrast bolus over the suspected leak site, while the patient, C-arm, or both are obliqued to optimize visualization of the ROI. Digital subtraction is not used in this technique. After a dynamic myelogram is obtained, the patient is transported to the CT scanner while maintaining the decubitus or prone position for the subsequent postmyelography CT scan, to maximize contrast concentration over the suspected leak site.

Image Analysis

All available spinal images for each patient were reviewed together by two board-certified radiologists who hold certificates of added qualification in neuroradiology and have 8 and 5 years of experience in treating patients with SIH, to determine whether a CSF venous fistula was present. Image review and analysis were performed using the thinnest available axial CTM images. If more than one CTM study was available to review, the study that best showed the fistula was identified and used for analysis. Any additional CTM studies performed on the same patient at other times or at other institutions were reviewed and designated as either showing or not showing the fistula. Spinal MRI examinations that may have been performed at outside institutions were not reviewed as part of this study because CSF venous fistulas are not detectable on MRI.
The spinal level and side where the CSF venous fistula was found were recorded. Images were also reviewed to ascertain whether a nerve root sleeve diverticulum was present at the origin of the fistula, whether there was any concurrent epidural leak of CSF, or whether any paraspinal vascular malformation was present.
For each CSF venous fistula identified, the primary location of the draining vein on CTM was characterized as either paravertebral (if it involved the spinal segmental vein adjacent to the vertebral body) (Fig. 1), lateral (if it involved an intercostal or muscular vein lateral to the neural foramen) (Fig. 2), or central (if it involved the internal vertebral epidural venous plexus) (Fig. 3). On CTM images, the presence of a hyperdense paraspinal vein sign was assessed. When the sign was present, the attenuation of the hyperattenuated vein was measured by drawing three oval ROIs over the vein on the axial images and recording the median value. The attenuation of veins at two adjacent levels not involved by the fistula was also measured using identical technique, to serve as a control. Control measurements were obtained first at the vertebral level below and ipsilateral to the hyperattenuated vein (designated as control 1) and, second, at two vertebral levels below and contralateral to the hyperattenuated vein (designated as control 2).
Fig. 1A —68-year-old man with positional headache caused by spontaneous intracranial hypotension.
A, Maximum-intensity-projection axial image from postmyelography CT shows contrast material filling left spinal segmental vein (arrow) in paravertebral location at T8 level. Scan was performed with patient in left lateral decubitus positon to maximize intrathecal contrast concentration on left side, because of suspected hyperdense paraspinal vein sign on prior CT myelogram (not shown).
Fig. 1B —68-year-old man with positional headache caused by spontaneous intracranial hypotension.
B, Anteroposterior projection of dynamic myelogram image from same patient, also obtained with patient in left lateral decubitus position, shows filling of same paraspinal vein at T8 (arrowhead).
Fig. 2A —41-year-old woman with severe positional headache of 3 months' duration. Brain MRI showed intracranial hypotension.
A, Maximum-intensity-projection axial (A) and coronal (B) images from postmyelography CT show contrast material filling muscular branch of vein lateral to spine at L1 (arrow), indicating presence of CSF venous fistula.
Fig. 2B —41-year-old woman with severe positional headache of 3 months' duration. Brain MRI showed intracranial hypotension.
B, Maximum-intensity-projection axial (A) and coronal (B) images from postmyelography CT show contrast material filling muscular branch of vein lateral to spine at L1 (arrow), indicating presence of CSF venous fistula.
Fig. 2C —41-year-old woman with severe positional headache of 3 months' duration. Brain MRI showed intracranial hypotension.
C, Oblique image from dynamic myelogram obtained at same level as images in A and B shows myelographic contrast material filling multiple small veins (arrowheads) in pattern replicating that seen on CT myelography.
Fig. 3A —59-year-old man with positional headaches of 2.5 months' duration and evidence of intracranial hypotension on brain MRI.
A, Axial image from postmyelography CT, obtained at T4 level, shows small focus of contrast material outside of thecal sac in internal vertebral epidural venous plexus (arrow).
Fig. 3B —59-year-old man with positional headaches of 2.5 months' duration and evidence of intracranial hypotension on brain MRI.
B, Anteroposterior projection of dynamic myelogram of same patient again shows contrast material filling epidural venous plexus (arrow) as well as opacification of collateral vessels inferior to nerve root sleeve and lateral to vertebral body (arrowheads).
Pretreatment brain MR images were reviewed to determine the presence of the principal signs of SIH, including dural enhancement, brain sagging, venous distention sign, and subdural collections [911].

Statistical Analysis

Mean attenuation values of the hyperdense paraspinal vein seen on CTM for all fistulas were compared with the attenuation values of the control veins at the adjacent level, with use of a two-sided t test. A p < 0.05 was considered to denote statistical significance.

Results

Demographic and Baseline Clinical Characteristics

A total of 22 cases of CSF venous fistula were identified. All subjects met the diagnostic criteria for SIH according to the International Classification of Headache Disorders, third edition [12]. Of these cases, 10 underwent surgical treatment, with the fistula confirmed in all cases. The subjects' demographic characteristics, CSF pressures, and brain imaging findings are presented in Table 1. Most of the subjects were women (73%).
TABLE 1: Demographic and Baseline Clinical Characteristics for 22 Subjects with CSF Venous Fistulas
CharacteristicValue
Age (y) 
 Mean (SD)53.0 (11.5)
 Range33-72
Sex 
 Female16 (73)
 Male6 (27)
HIT-6 score 
 Mean (SD)68.3 (4.8)
 Range62-78
MRI finding 
 Any sign of SIH20 (91)
 Dural enhancement16 (73)
 Brain sagging14 (64)
 Venous distention sign17(77)
 Subdural collection4 (18)
Opening pressure (cm H2O) 
 Mean (SD)7.8 (4.4)
 Range0-17.5

Note—Except where otherwise indicated, data are number (%) of patients. HIT-6 = Headache Impact Test, SIH = spontaneous intracranial hypotension.

HIT-6 scores at baseline were available for 16 of 22 subjects. For all these subjects, the baseline HIT-6 score was greater than 60, corresponding to very severe headache intensity. A positional component to the headaches was present in all patients
CSF opening pressures ranged from 0 to 17.5 cm H2O, with only 32% of patients having an opening pressure of less than 6 cm H2O. Pretreatment brain MR images were available in all cases. The pretreatment MR images were obtained with contrast in all but one case. At least one sign of SIH was present in 91% of cases.

Anatomic Features

Fistulas were located between T4 and L1, with most fistulas located in the lower half of the thoracic spine at T7–T12 (68%). No fistulas were identified in the cervical spine or below L1. Fifteen fistulas were located on the left side of the spinal column and seven on the right.
Fistulas were associated with a nerve root sleeve diverticulum in 82% of cases (18/22) (Fig. 4). In most cases, no concurrent epidural CSF leak was present, with leaks seen in only 9% of cases (2/22).
Fig. 4A —72-year-old woman with positional headache of 8 months' duration and evidence of spontaneous intracranial hypotension on brain MRI.
A, Maximum-intensity-projection (MIP) axial image shows CSF venous fistula originating from nerve root sleeve diverticulum at T7 level (arrow). Contrast material opacifies paravertebral vein as well as collateral veins within vertebral body (arrowheads).
Fig. 4B —72-year-old woman with positional headache of 8 months' duration and evidence of spontaneous intracranial hypotension on brain MRI.
B, MIP sagittal image from same scan again shows diverticulum (arrow) at origin of fistula and filling of draining paravertebral vein (white arrowhead). Note much lower attenuation of veins at adjacent levels not involved by fistula (black arrowheads).
Fig. 4C —72-year-old woman with positional headache of 8 months' duration and evidence of spontaneous intracranial hypotension on brain MRI.
C, Oblique image obtained using dynamic myelography again shows fistula originating from diverticulum (arrow) and draining into paravertebral vein (arrowhead).
Paraspinal vascular malformations were present in three cases in this series. These included two venous malformations and one venolymphatic malformation. In all three cases, these malformations were located in the paraspinal soft tissues on the same side and at the same level as the fistula (Fig. 5).
Fig. 5A —57-year-old with 8-year history of headache worsened by upright posture and Valsalva maneuver.
A, Anteroposterior projection from dynamic myelography shows very irregular appearance of left T5 nerve root (arrow) with irregular contrast opacification of network of channels inferior to nerve root (arrowhead).
Fig. 5B —57-year-old with 8-year history of headache worsened by upright posture and Valsalva maneuver.
B, Axial T1-weighted contrast-enhanced MR image of thoracic spine shows diffuse venous malformation involving paravertebral soft tissues (arrowheads) and extending up to neural foramen at T5 (arrow), where CSF venous fistula was seen on dynamic myelography.

Imaging Features on Dynamic Myelography and CT Myelography

Seventy-three percent of subjects (16/22) underwent both CTM and dynamic myelography. Of these, 14/16 (88%) had fistulas seen on both imaging modalities. In one case (6%), the fistula was seen on dynamic myelography but not on CTM, and in one case (6%), it was seen on CTM but not on dynamic myelography. For the remaining six subjects who did not undergo both forms of imaging, CTM showed a clear CSF venous fistula, and the treating physician thought that confirmatory dynamic myelography was not needed.
For 15 of 22 subjects, more than one CTM study had been performed for the subject during the course of their workup. For 10 of these 15 subjects, at least one of the additional CTM studies failed to identify the fistula (Fig. 6). In most cases (12/15) for which CTM was performed more than once, the fistula was best identified on CTM performed at our institution.
Fig. 6A —52-year-old-woman with 11-month history of headache that worsened with bending over and Valsalva maneuver. Brain MRI showed intracranial hypotension.
A, Axial image from CT performed after myelography shows hyperdense paraspinal vein sign (arrowheads) because of CSF venous fistula at T7 level.
Fig. 6B —52-year-old-woman with 11-month history of headache that worsened with bending over and Valsalva maneuver. Brain MRI showed intracranial hypotension.
B, Anteroposterior projection from dynamic myelography shows CSF venous fistula (arrowheads) arising from nerve root sleeve diverticulum at this level.
Fig. 6C —52-year-old-woman with 11-month history of headache that worsened with bending over and Valsalva maneuver. Brain MRI showed intracranial hypotension.
C, Axial image from CT performed after myelography at outside institution does not show hyperattenuated vein. Image shown is from same level as image shown in A and is displayed using same window and level settings. Note that concentration of contrast material in thecal sac is lower in study that did not show fistula.
The attenuation values of the hyperattenuated paraspinal veins marking the CSF venous fistulas and the attenuation values of the control veins are shown in Fig. 7. The median attenuation value of the hyperattenuated paraspinal vein identified on CTM was 144 HU (range, 77–781 HU). The median attenuation value was 36 HU (range, 10–63 HU) for the control 1 veins and 34 HU (range, 4–50 HU) for the control 2 veins. The differences in attenuation between the fistula veins and each of the control vein sets was highly statistically significant (p < 0.0001). A threshold value of 70 HU perfectly discriminated the fistula veins from the control veins in the present study.
Fig. 7 —Comparison of attenuation of hyperattenuated paraspinal veins making CSF venous fistulas with attenuation of adjacent level control veins on CT myelography. Control 1 denotes attenuation measurements obtained at vertebral level below and ipsilateral to hyperattenuated vein, and control 2 denotes attenuation measurements obtained at two vertebral levels below and contralateral to hyperattenuated vein.
A paravertebral location was the most common site of a hyperattenuated paraspinal vein (45%), with a central location accounting for 32% of all sites and a lateral location accounting for 23% of all sites. In the single case in which the fistula was not identified on dynamic myelography, the vein was in a paravertebral location and was obscured by an over-lying nerve root sleeve diverticulum on the anteroposterior myelographic projection.

Treatment Results

All subjects underwent attempted targeted epidural patching using a combination of autologous blood and fibrin glue (Tisseel, Baxter) at the level of the fistula as first-line treatment. For three cases, targeted epidural patching resulted in resolution of SIH symptoms. For nine cases, the treatment failed or the subject had only partial improvement after initial patching; several of these subjects currently have surgery scheduled. For 10 subjects, epidural patching failed and surgical treatment of the fistula was performed; all these subjects had immediate resolution of their symptoms related to CSF leak.

Discussion

CSF venous fistula is a recently recognized cause of SIH that can occur in the absence of a detectable epidural CSF leak. It is well recognized that a relatively large percentage of cases of SIH may show no epidural CSF leak despite exhaustive imaging, limiting treatment options if nontargeted blood patching fails [4, 13, 14]. Recognition of CSF venous fistulas is therefore important because it may explain some of these cases found to have negative myelography results and also provides a target for treatment. Experience with this entity currently is quite limited, however, with fewer than 20 total cases reported in the literature; therefore, most radiologists will be unfamiliar with the location and appearance of CSF venous fistulas [13, 5, 15].
Our investigation highlights the anatomic features of CSF venous fistulas that may aid in their detection. In particular, CSF venous fistulas seem to occur in the thoracic spine and at the thoracolumbar junction, and they often occur without a concurrent epidural CSF leak. CSF venous fistulas are most frequently paravertebral in location, but they can also occur in other locations that are lateral to or even within the spinal canal. Careful scrutiny of adjacent veins, including the epidural venous plexus, is important; in some cases, fistulas in a central location may be mistaken for an epidural leak. Paraspinal vascular malformations were present in three of our subjects, suggesting that patients with SIH who also have vascular malformations should undergo careful evaluation for an associated CSF venous fistula.
It is interesting to note that a high percentage of CSF venous fistulas in our series (82%) were associated with a nerve root sleeve diverticulum. Nerve root sleeve diverticula have been a controversial topic in SIH: diverticula are often found at the site of spontaneous CSF leaks, but they also occur in patients without SIH, where they presumably are incidental findings of no consequence [16]. With regard to CSF venous fistulas, however, this high rate of association between diverticula and CSF venous fistulas may imply a pathophysiologic connection. For example, it has been suggested that some diverticula may represent enlarged arachnoid granulations [8]. Arachnoid granulations are known to be present in the spine and are intimately associated with paraspinal veins [17]. Rupture of an enlarged arachnoid granulation into its adjacent vein could be one possible explanation or the development of fistulas, and it might explain the association that we observed between CSF venous fistulas and diverticula.
The present study also suggests some important insights regarding the detection of CSF venous fistulas with the use of CTM and dynamic myelography. First, most cases were seen with both dynamic myelography and CTM, although each modality missed one CSF venous fistula that was detected by the other modality. Therefore, in some cases, it may be necessary to perform more than one imaging test to detect a CSF venous fistula. Because these lesions may be subtle, and because the area of venous contrast opacification is small, thin-section imaging (we use an axial slice thickness of 0.625 mm for analysis) is crucial.
Second, in a substantial number of the cases in which a CSF venous fistula was identified, at least one additional CTM had been performed that did not show the fistula. It is unclear whether this is caused by differences in technique between different scanning methods or patient-specific physiologic variables; because a greater number of CSF venous fistulas are recognized, it will be important to examine what specific imaging protocols best facilitate their detection.
Third, a threshold of 70 HU was a reliable cutoff value for distinguishing the site of fistulas from normal veins in our series. Identifying a threshold that defines a vein as hyperattenuated is important to increase diagnostic confidence that a CSF venous fistula is present, although in some cases, confirmation with dynamic myelography or DSM may still be desirable, particularly if surgery is planned. A prior report on CSF venous fistulas has suggested that image analysis of CTMs using automated thresholding makes some fistulas more conspicuous, a process that also requires input of such a threshold value [5]. It is possible that the exact threshold may vary within institutions because of technical differences.
Clearly, it would be advantageous to define a single imaging test or protocol that is highly sensitive for detecting both CSF venous fistulas and other types of CSF leaks. Although CTM certainly has advantages over dynamic myelography in the initial workup when the cause of SIH is unknown, and although it performs well in comparison with dynamic myelography in the detection of CSF venous fistulas overall, we found that dynamic myelography was still useful for confirming CSF venous fistulas and discriminating some central CSF venous fistulas from low-flow epidural leaks. In practice, we use CTM as an initial screening test for patients with SIH, and dynamic myelography as a confirmatory or problem-solving test when necessary. Although MRI is performed at some institutions as an initial screening test for spinal epidural CSF leaks, it has not yet been shown to be capable of detecting CSF venous fistulas.
Optimal treatment of CSF venous fistulas is not yet defined. Long-term outcome data for subjects in the present study are not currently available; however, some preliminary observations may be useful in choosing treatment for patients with CSF venous fistulas. Most patients in our series did not completely respond to epidural patching, with most ultimately requiring surgery. Nevertheless, three of our 22 of patients did respond to targeted epidural patching. We speculate that the mechanism underlying this response to patching may be that mass effect from the injectate temporarily compresses the draining veins long enough to allow healing of the underlying fistulous connection. Because epidural patching is less invasive than surgery and requires shorter recovery time, a trial of patching for patients who wish to avoid surgery is reasonable to consider.
It is important to recognize that this investigation likely is affected by detection bias, in that CTM is the first-line spinal imaging modality used at our institution and therefore the CSF venous fistulas that we detected likely are those most readily seen on CTM. Our investigation does not address the sensitivity of CTM for CSF venous fistula detection, because there is no current reference standard that is known to reliably detect all (or a high percentage) of CSF venous fistulas, and because the true prevalence of this lesion in patients with SIH is not currently known. Furthermore, the present study did not address the use of DSM, and therefore, on the basis of our data, no conclusions can be drawn regarding the comparative sensitivity of our dynamic myelographic technique or that of CTM to DSM.
It is notable, however, that temporal resolution—a benefit of DSM—seemed not to be a critical factor in detecting many CSF venous fistulas, because fistulas were detected even on CT performed at outside institutions, where imaging did not occur until 20–30 minutes after contrast injection. In our opinion, achieving a high concentration of myelographic contrast material at the level of the fistula and finding ways to augment flow though the fistula to promote passage of contrast material into the vein are likely to be more important than temporal resolution in the detection of these lesions, at least up to a point; excessive delays in imaging after contrast injection would likely result in reduced detection of the fistula. It is possible that, even after setting aside the issue of temporal resolution, DSM may make subtle fistulas more conspicuous on myelography because of the subtraction of background structures, and therefore it may prove valuable compared with unsubtracted myelography for some fistulas.
In conclusion, CSF venous fistulas are an important and recently discovered type of lesion that can cause SIH, and they are more frequently being recognized in patients with no other identifiable epidural leak. Familiarity with the anatomic appearance and imaging characteristics of these lesions is critical, because recognition of CSF venous fistulas may provide treatment options to patients with SIH who have been otherwise refractory to therapy.

References

1.
Schievink WI, Moser FG, Maya MM. CSF-venous fistula in spontaneous intracranial hypotension. Neurology 2014; 83:472–473
2.
Schievink WI, Moser FG, Maya MM, Prasad RS. Digital subtraction myelography for the identification of spontaneous spinal CSF-venous fistulas. J Neurosurg Spine 2016; 24:960–964
3.
Kumar N, Diehn FE, Carr CM, et al. Spinal CSF venous fistula: a treatable etiology for CSF leaks in craniospinal hypovolemia. Neurology 2016; 86:2310–2312
4.
Kranz PG, Luetmer PH, Diehn FE, Amrhein TJ, Tanpitukpongse TP, Gray L. Myelographic techniques for the detection of spinal CSF leaks in spontaneous intracranial hypotension. AJR 2016; 206:8–19
5.
Kranz PG, Amrhein TJ, Schievink WI, Karikari IO, Gray L. The “hyperdense paraspinal vein” sign: a marker of CSF-venous fistula. AJNR 2016; 37:1379–1381
6.
Rendas-Baum R, Yang M, Varon SF, Bloudek LM, DeGryse RE, Kosinski M. Validation of the Headache Impact Test (HIT-6) in patients with chronic migraine. Health Qual Life Outcomes 2014; 12:117
7.
Yang M, Rendas-Baum R, Varon SF, Kosinski M. Validation of the Headache Impact Test (HIT-6) across episodic and chronic migraine. Cephalalgia 2011; 31:357–367
8.
Kranz PG, Gray L, Taylor JN. CT-guided epidural blood patching of directly observed or potential leak sites for the targeted treatment of spontaneous intracranial hypotension. AJNR 2011; 32:832–838
9.
Kranz PG, Tanpitukpongse TP, Choudhury KR, Amrhein TJ, Gray L. Imaging signs in spontaneous intracranial hypotension: prevalence and relationship to CSF pressure. AJNR 2016; 37:1374–1378
10.
Farb RI, Forghani R, Lee SK, Mikulis DJ, Agid R. The venous distension sign: a diagnostic sign of intracranial hypotension at MR imaging of the brain. AJNR 2007; 28:1489–1493
11.
Fishman RA, Dillon WP. Dural enhancement and cerebral displacement secondary to intracranial hypotension. Neurology 1993; 43:609–611
12.
Headache Classification Committee of the International Headache Society (IHS). The International Classification of Headache Disorders, 3rd edition (beta version). Cephalalgia 2013; 33:629–808
13.
Sencakova D, Mokri B, McClelland RL. The efficacy of epidural blood patch in spontaneous CSF leaks. Neurology 2001; 57:1921–1923
14.
Luetmer PH, Schwartz KM, Eckel LJ, Hunt CH, Carter RE, Diehn FE. When should I do dynamic CT myelography? Predicting fast spinal CSF leaks in patients with spontaneous intracranial hypotension. AJNR 2012; 33:690–694
15.
Schievink WI, Maya MM, Jean-Pierre S, Nuño M, Prasad RS, Moser FG. A classification system of spontaneous spinal CSF leaks. Neurology 2016; 87:673–679
16.
Kranz PG, Stinnett SS, Huang KT, Gray L. Spinal meningeal diverticula in spontaneous intracranial hypotension: analysis of prevalence and myelographic appearance. AJNR 2013; 34:1284–1289
17.
Pollay M. The function and structure of the cerebrospinal fluid outflow system. Cerebrospinal Fluid Res 2010; 7:9

Information & Authors

Information

Published In

American Journal of Roentgenology
Pages: 1360 - 1366
PubMed: 29023155

History

Submitted: April 5, 2017
Accepted: April 22, 2017
Version of record online: October 12, 2017

Keywords

  1. CSF hypovolemia
  2. CSF leak
  3. CSF venous fistula
  4. myelography
  5. spontaneous intracranial hypertension

Authors

Affiliations

Peter G. Kranz
All authors: Department of Radiology, Duke University Medical Center, Box 3808, Durham, NC 27710.
Timothy J. Amrhein
All authors: Department of Radiology, Duke University Medical Center, Box 3808, Durham, NC 27710.
Linda Gray
All authors: Department of Radiology, Duke University Medical Center, Box 3808, Durham, NC 27710.

Notes

Address correspondence to P. G. Kranz ([email protected]).

Metrics & Citations

Metrics

Citations

Export Citations

To download the citation to this article, select your reference manager software.

Articles citing this article

View Options

View options

PDF

View PDF

PDF Download

Download PDF

Media

Figures

Other

Tables

Share

Share

Copy the content Link

Share on social media