Secondary Logo

Journal Logo

High Thoracic/Cervical Epidural Blood Patch for Spontaneous Cerebrospinal Fluid Leak: A New Challenge for Anesthesiologists

Nipatcharoen, Panya MBBS, FANZCA; Tan, Suyin G. MBBS, FRCA, FANZCA, FFPMANZCA

doi: 10.1213/ANE.0b013e31822e5665
Neuroscience in Anesthesiology and Perioperative Medicine: Case Report

Spontaneous cerebral spinal fluid leakage is increasingly recognized as a cause of headache due to low intracranial pressure. The site of leakage can be identified with radionuclide cisternography, and anesthesiologists are increasingly requested to provide epidural blood patch for their management. This series of case reports demonstrates some of the issues relating to the management of this condition.

Published ahead of print September 29, 2011 Supplemental Digital Content is available in the text.

From the Department of Anaesthetic and Pain Management, Nepean Hospital, Penrith, New South Wales, Australia.

Funding: None.

The authors declare no conflict of interest.

Reprints will not be available from the authors.

Address correspondence to Dr. Panya Nipatcharoen, Department of Anaesthesia and Pain Management, Nepean Hospital, Derby St., Kingswood, NSW 2747 Australia. Address e-mail to

Accepted July 6, 2011

Published ahead of print September 29, 2011

Patients presenting with subacute or chronic orthostatic headache are increasingly recognized as having spontaneous dural tears resulting in cerebral spinal fluid (CSF) leakage and subsequent low CSF pressure. The dural tear often occurs at the thoracic or cervical level.1 There is some evidence that epidural blood patch (EBP) at the site of leakage is more effective than EBP at a different level.2 However, most anesthesiologists are unfamiliar with performing cervical or thoracic level EBP, and there is increased risk involved. Conversely, there are a few reports suggesting that lumbar EBP, performed without localization of the site of leakage, may provide complete or partial resolution of the headache without the risks associated with a higher-level EBP.1,3 Imaging techniques are useful in both identifying the site of leakage and placement of EBP and may facilitate safer and more effective treatment of this condition.

In this case series, we report that lumbar EBP failed to provide long-lasting relief of the headache when the site of CSF leakage was identified to be in the cervical or high-thoracic region. We also describe the use fluoroscopy in identifying the spinal level and placement of the Tuohy needle into the epidural space.

Back to Top | Article Outline


A 27-year-old woman presented initially with a 2-week history of severe, constant bilateral, frontal-temporal headache. The headache had a gradual onset, with intermittent increases in intensity. It was associated with nausea and vomiting and exacerbated by upright posture, coughing, and straining. The general practitioner diagnosed her with headache secondary to uncontrolled hypertension (her systolic blood pressure was >200 mm Hg on presentation). She was admitted to hospital initially under the care of a renal physician to control her hypertension. Her initial computed tomography (CT) scan and renal ultrasound were normal. She was discharged home with perindopril and atenolol to control her hypertension and amitryptylline for her headache.

Despite the above treatment, the headache remained severe. Her headache was worse on standing. She presented again 2 weeks later, and a magnetic resonance imaging (MRI) scan revealed marginal leptomeningeal enhancement with no other abnormalities detected, suggesting a low CSF pressure headache. Her radioisotope cisternography (RC) study identified a leakage of CSF in the low cervical region.

EBP at the level of T12-L1 was performed (as requested by the neurologist), during which 12 mL of autologous blood was injected into the epidural space. She remained in the Trendelenburg position for about 2 hours. She experienced some improvement for approximately 24 to 48 hours, after which the severity of the headache returned to the level before the procedure. A second EBP patch was performed under fluoroscopy guidance at the level of C6-C7 at which 8 mL of autologous blood was injected. Her headache improved immediately and completely disappeared within 2 to 3 days. At 6 weeks follow-up, she reported return to normal activities.

Back to Top | Article Outline


A 22-year-old woman presented to the emergency department multiple times with a persistent headache. Her initial presentation was thought to be a result of viral upper respiratory tract infection. She returned 4 weeks later with frontal headache, which had a postural component. There were no other symptoms. Her lumbar puncture was normal. Low CSF pressure headache was diagnosed, and lumbar EBP was performed (without RC study to determine the level). Her symptoms improved substantially, and a subsequent MRI scan was consistent with low-pressure headache. The patient was discharged with mild symptoms. She was readmitted 5 days later owing to the return and worsening of her headache. RC study confirmed CSF leakage at the low cervical level.

EBP was performed at the level of C6-C7 under fluoroscopy guidance. Eight milliliters of autologous blood was injected into the space without any difficulties. She went home after the procedure with markedly improved symptoms. She was reviewed in the clinic 3 weeks later and her headache had completely resolved.

Back to Top | Article Outline


A 44-year-old woman presented to the hospital with 1-week history of retrobulbar headache worse when she sat or stood up. The patient reported that she had been straining because of constipation for a few weeks before her presentation. Her CT brain scan was normal. Lumbar puncture was attempted, but there was no flow of CSF. MRI scan showed dural enhancement, and RC study revealed parathecal tracer activity at the cervicothoracic junction and possibly a small leakage at the T12 level.

EBP was performed under fluoroscopy guidance at the level of T2. Fifteen milliliters of autologous blood was injected into the epidural space. The patient had symptomatic relief for about 12 hours, at which time her symptoms returned. The EBP was repeated at the same level a few days later. Her symptoms improved and within the next few days she was discharged with a mild headache. Her headache had completely resolved when she was reviewed in the clinic 4 weeks later.

Back to Top | Article Outline


A 30-year-old man presented with 1-week history of retrobulbar throbbing headache with exacerbation on standing and associated with nausea, vomiting, and diplopia. There was no fever, neck stiffness, or other neurological symptoms. A brain CT scan and lumbar puncture revealed no abnormality. Low CSF pressure headache was suspected. The RC revealed bilateral CSF leakage around T5-T7. EBP was performed at the level of T5 with the aid of fluoroscopy, and 18 mL of autologous blood was injected into the epidural space. His symptoms immediately improved and had resolved the next day. He was discharged the following day. There were no further symptoms at 6 months.

Back to Top | Article Outline


Low CSF pressure headache is almost always due to CSF leakage, which can be iatrogenic, traumatic, or spontaneous in origin. The etiology of spontaneous CSF leakage remains unclear, and the leakage site can be anywhere along the spinal column but is usually at the thoracic or low cervical level.4 The most common presentation is orthostatic headache. Other features may include cranial nerve palsy, dizziness, tinnitus, photophobia, change in hearing, or neck stiffness.

A number of investigations can be performed to confirm the diagnosis, including CSF examination, CT, and MRI scans. CSF examination may confirm a low CSF pressure but does not indicate the level of the leak. In most cases, CT scan has little value because the result is usually normal. MRI scan may demonstrate diffuse pachymeningeal enhancement, which is a classical feature of low CSF pressure, but often fails to identify the site of CSF leakage.

With RC, the CSF leakage site can be identified and the diagnosis can be confirmed, although its sensitivity is low.5 This is due to the low resolution of the study and the low rate of leakage. One hundred megabequerel units (2.5 mCi) of radionuclide (Technetium) is injected intrathecally using a 27-G or 29-G spinal needle in the lumbar spine under fluoroscopy or CT guidance. Immediately after the intrathecal injection, the patient is placed in the supine position, and an initial cisternography is performed. The nuclear scan is repeated at 30-minute intervals to a maximum of 24 hours or until a definite trace leak out of the CSF compartment is detected. This denotes the spinal level of CSF leak.

The most common cisternographic abnormality in CSF leaks is that the radioactivity typically does not extend much beyond the basal cisterns, and at 24 and even 48 hours after injection, there is an absence or paucity of activity over the cerebral convexities. A much less common abnormality is the detection of parathecal activity, indicating the approximate site of the leak (Fig. 1). In our series, RC confirmed the site of leakage, and EBP was performed at these levels.

Figure 1

Figure 1

In anesthetic practice, EBP at the leakage site is the treatment of choice if conservative measures have failed. Other treatments have been suggested with different success rates, including intrathecal fluid infusion, epidural saline infusion, epidural infusion of dextran, epidural injection of fibrin glue, CSF shunting, and surgical repair of the leak.6 In general, EBP can be safely performed at the lumbar level, but in this series, the site of CSF leakage had been identified at the lower cervical level or at the cervicothoracic junction. The procedure then becomes more technically challenging and higher risk.

There are several reports indicating that lumbar EBP can permanently alleviate the headache regardless of whether or not the site of leakage is identified.3,7,8 However, other reports demonstrate that lumbar EBP does not always result in permanent relief.4,9,10 A study by Diaz11 suggests that the site of leakage should be identified by RC and treated with EBP targeted to CSF leak site levels. A report by Kantor and Silberstein12 also suggests that cervical EBP may be useful after failed lumbar EBP when the site of leakage is unidentified.

A case report by Ferrante et al.3 indicated that lumbar EBP successfully treated spontaneous CSF leak at the cervical region. This report also showed the spread of the blood from the lumbar to cervical region.

In this case series, 2 patients received lumbar EBP initially. Both had transient symptomatic relief; however, both subsequently required further EBP at the level of CSF leakage as identified by RC study. They both experienced immediate symptomatic relief and were completely free of symptoms during the subsequent follow-up. Our 3rd patient did not receive an initial lumbar EBP, but she required repeated EBP at T2 level, at which the leakage was identified. We believe that the reason for the failed initial EBP was in part due to the small leakage at the T12 level. However, the decision was made to perform a repeat EBP at the higher level because this was the main site of leakage demonstrated on RC. Because her headache improved after the repeat EBP, it was not necessary to perform another EBP at the T12 level. The 4th patient received 1 EBP at the suspected leakage site, and he was completely free of symptoms within 24 hours.

Cousins et al.2 suggested that placement of the EBP close to the site of CSF leakage is important. The proposed mechanism is that the injected blood seals the dural defect and stops the leakage. The other theory is that the injected blood causes an epidural tamponade effect over the leakage site. It would seem sensible to target the treatment at the site of the leakage to maximize the chances of success, but there is no clear evidence to support targeted EBP, and randomized clinical trials are unlikely to be feasible given the low incidence of the condition.

Most anesthesiologists are familiar with performing EBP using palpation to identify the bony landmarks and hence spinal level. However, this technique is frequently inaccurate, particularly in obese or scoliotic patients. The use of fluoroscopy allows accurate identification of the spinal level and confirmation of epidural injection of the blood.4,13 The EBP is performed under aseptic conditions in the operating room. The patient is positioned prone on the operating table, and fluoroscopy is used to identify the correct spinal level and angle of approach (Fig. 2). A 16-G Touhy needle is inserted epidurally by standard technique using loss-of-resistance to saline. The depth of the epidural space is not routinely confirmed with a lateral view on fluoroscopy because there are often technical issues with obtaining a good lateral view at the cervical and upper thoracic level because of the position of the shoulders. Entry into the epidural space is confirmed by injecting 2 to 3 mL of water-soluble radio-contrast. This indicates the potential extent of spread, confirms accurate placement, and avoids inadvertent intravascular or subarachnoid injection.14 Autologous blood is collected by a 2nd medical practitioner using an aseptic technique. Blood is injected slowly until the patient indicates that he or she has discomfort in the site. After EBP, the patient is advised to lie supine for 1 to 2 hours. Postprocedure standard observations include cardiovascular, respiratory, and extremity neurological assessment.

Figure 2

Figure 2

The use of fluoroscopy is more convenient than using a CT scan–guided technique because it is more readily available. A study by Bhandari et al.15 found that using fluoroscopic guidance for EBP resulted in better outcomes for postdural puncture headache, although they suggested that this technique be reserved for patients with altered anatomic landmarks (e.g., with previous lumbar spine surgery) or those who did not have adequate relief of symptoms with conventional (blind) technique.

There is controversy over whether low-pressure headache should be managed with simple lumbar EBP or with technically more challenging localization of CSF leakage and EBP at the identified level. The latter is more time and resource consuming but has the potential to provide more effective treatment for patients.

Back to Top | Article Outline


Low CSF pressure headache due to spontaneous leakage has become increasingly recognized by neurologists resulting in increased referral to anesthesiologists for EBP. The availability of RC imaging clarifies the diagnosis and allows identification of the leakage site. This case series supports the view that performing EBP at the site of suspected leakage is more effective. The use of fluoroscopy to assist in identifying the spinal level can be useful especially when performing EBP at cervical or thoracic levels, and this may reduce the risk of serious complications.

Back to Top | Article Outline


Name: Panya Nipatcharoen, MBBS, FANZCA.

Contribution: Collecting data, literature review, and main author.


Contribution: Original idea, manuscript editing.

This manuscript was handled by: Gregory J. Crosby, MD.

Back to Top | Article Outline


1. Schievink W, Meyer F, Atkinson J, Mokri B. Spontaneous spinal cerebrospinal fluid leaks and intracranial hypotension. J Neurosurg 1996;84:598–605
2. Cousins MJ, Brazier D, Cook R. Intracranial hypotension caused by cervical cerebrospinal fluid leak: treatment with epidural blood patch. Anesth Anal 2004;98:1794–7
3. Ferrante E, Arpino I, Citterio A. Is it a rational choice to treat with lumbar epidural bloodpatch headache caused by spontaneous cervical CSF leak? Cephalalgia 2006;26:1245–6
4. Terjima K, Oi Y, Ogura A, Sakai N, Takei M, Tamaki T, Ogawa R. Blood patch therapy for spontaneous intracranial hypotension: safe performance after epidurography in an unconscious patient. Anesth Anal 2002;94:959–61
5. Horikoshi T, Ikegawa H, Uchida M, Takahashi T, Watanabe A, Umeda T. Tracer clearance in radionuclide cisternography in patients with spontaneous intracranial hypotension. Cephalagia 2006;26:1010–5
6. Mokri B. Headaches caused by decreased intracranial pressure: diagnosis and management. Curr Opin Neurol 2003;16:319–26
7. Benzon HT, Nemickas R, Molloy RE, Ahmad S, Melen O, Cohen B. Lumbar and thoracic epidural blood injections to treat spontaneous intracranial hypotension. Anesthesiology 1996;85:920–2
8. Berroir S, Loisel B, Ducros A, Boukobza M, Tzourio C, Valade D, Bousser M-G. Early epidural blood patch in spontaneous intracranial hypotension. Neurology 2004;63:1950–1
9. Hannerz J, Dahlgren G, Irestedt L, Meyerson B, Ericson K. Treatment of idiopathic intracranial hypotension: cervicothoracic and lumbar blood patch and peroral steroid treatment. Headache 2006;46:508–11
10. Nowak DA, Takano B, Topka H. Spontaneous cerebrospinal fluid hypovolaemia: a therapeutic dilemma? Eur J Neurol 2006;13:1128–38
11. Diaz JH. Epidemiology and outcome of postural headache management in spontaneous intracranial hypotension. Reg Anesth Pain Med 2001;26:582–7
12. Kantor D, Silberstein SD. Cervical epidural blood patch for low CSF pressure headaches. Neurology 2005;65:1138
13. Rai A, Rosen C, Carpenter J, Miele V. Epidural blood patch at C2: diagnosis and treatment of spontaneous intracranial hypotension. Am J Neuroradiol 2005;26:2663–6
14. Waguri N, Tomita M, Hayatsu K, Okamoto K, Shimoji K. Epidural blood patch for treatment of spontaneous intracranial hypotension. Acta Anaesthesiol Scand 2002;46:747–50
15. Bhandari A, Anand A, Khan F, Ellias M. Epidural blood patch: comparison of fluoroscopic guided technique versus non-fluorsoscopic (blind) technique. The Pain Clinic 2001;13:77–82
© 2011 International Anesthesia Research Society