_{Intracranial hypotension headache after uncomplicated caudal epidural injection}

Author Information
Anaesthetic Research Fellow, Pain and Anaesthesia Research Centre, St. Bartholomew’s Hospital, London, UK.  Consultant in Anaesthesia and Pain Medicine, Broomfield Hospital, Chelmsford, UK

Summary
A caudal epidural injection was performed on a middle-aged woman for pain in her right foot. Although the procedure was uncomplicated and a good epidurogram was obtained, the patient went on to develop an orthostatic headache with generalised weakness and syncopal episodes that were treated successfully by epidural blood patching. We describe the aetiology, presentation and treatment of spontaneous intracranial hypotension and review the similarities with our patient’s clinical presentation. We hypothesise as to how our intervention may have resulted in a dural tear.

The most common form of iatrogenic intracranial hypotension occurs secondary to accidental or deliberate dural puncture with a needle. Iatrogenic intracranial hypotension of this nature produces the well-characterised syndrome of postdural puncture headache first documented by Bier. As a result of advances in imaging modalities, intracranial hypotension due to spontaneous cerebrospinal fluid leak has been recognised as a cause of headache in the last 20 years, with an estimated prevalence of 1 in 50 000. Whilst intracranial hypotension secondary to spontaneous cerebrospinal fluid leak includes orthostatic headache as one of its cardinal features, a wide variety of headache patterns and other symptoms are known to present, including stupor and coma. Although termed spontaneous, a history of trivial trauma can be elicited in about one third of patients. We describe a case of orthostatic headache in the setting of a recent caudal epidural injection.

Case report
A middle-aged female patient presented to the pain clinic with lower back and radicular pain in the right leg, made worse by sitting and with occasional sleep disturbance. In the preceding year she had undergone treatment by a chiropractor and an acupuncturist, neither of which had been of any benefit. Her current medication was ibuprofen as required and nortriptylline at night. After uncomplicated pulsed radio frequency treatment of the L5 and S1 dorsal root ganglion her back pain improved, but she continued to experience troublesome pain in the right foot and she was scheduled for a caudal epidural injection.


After infiltration with local anaesthetic and without sedation, a caudal epidural was performed with a 50-mm 21-G needle under image intensifier guidance. A good epidurogram was demonstrated and a mixture of 80 mg depomedrone and 10 ml bupivicaine 0.25% was injected into the epidural space. The procedure was performed easily with no immediate complications noted at the time. Twelve days later, the patient presented with severe postural headache to the point where she could no longer sit or stand. On examination, apart from photophobia, there were no abnormal neurological signs and Kernig’s sign was negative. Urgent magnetic resonance imaging (MRI) of the brain and spine was normal except for an annular disc bulge at L4/5 and a small annular disc bulge at the lower cervical discs; the dural sac terminated at L4. After some improvement with simple analgesics she was discharged home with advice to continue regular analgesia and to drink a minimum of 4 l fluid per day. The patient was followed up over the telephone three days later. She was still complaining of postural headache, which was now associated with generalised body pain and weakness. She felt that she was coping and declined further evaluation. Over the following two weeks, the headache did not improve and her symptoms began to evolve and at this point, she was persuaded to attend hospital.

On admission, the patient reported spending most of her time recumbent, rising periodically only to complete basic self-care; she had, for all intents and purposes, become housebound. She had a constant headache and had developed syncopal episodes on sitting for longer than 5 min. Routine monitoring established that these episodes were not accompanied by any cardiovascular disturbance. They were, however, followed by a dissociative state lasting up to 15 min. On regaining consciousness the patient described a state of awareness without the ability to speak or follow motor commands. The headache was now associated with a buzzing sensation in the face, intermittent visual disturbance and an aversion to noise. When supine and conscious, neurological examination was entirely normal. Clinical examination of other body systems was unremarkable and routine blood tests revealed no abnormalities. The MRI of her brain and spine was repeated and reported as normal with the dural sac confirmed as terminating at L4.

Despite the uncomplicated initial procedure and the proven spinal anatomy, the patient was treated as having postdural puncture headache. Initial treatment comprised copious intravenous rehydration, ketoprofen, diazepam and amitriptylline. During the patient’s admission, daytime somnolence became a continuous feature and she was seen by a neurologist. A subsequent electroencephalogram was normal and low intracranial pressure headache was diagnosed. Epidural blood patching was performed under fluoroscopic guidance at the L5/S1 interspace. Thirty millilitres of autologous blood was injected into the epidural space without pain or immediate complication. The patient’s symptoms resolved within 2 h, and after further observation she was discharged home. She remained symptom-free for two months and then began to experience further occipital headaches and buzzing sensations after standing for more than 2 h. Computed tomographic (CT) myelography and enhanced MRI were performed, but did not demonstrate a cerebrospinal fluid leak. The epidural blood patch was repeated through the caudal space and was again immediately successful, enabling the patient to return to work. At 6-week follow-up, she was able to stand for more than 4 h continuously, experiencing tolerable pain in the greater occipital nerve distribution only.

Discussion
Although the temporal relationship between the caudal injection and the onset of the patient’s symptoms is strong, other factors argue against this mechanism of injury. The procedure was performed with a 50-mm needle through the sacral hiatus and an MRI of the patient’s spine confirmed the dural sac to terminate at L4. A conservative estimate would suggest a 10-cm gap between the tip of the needle and the end of the dura. To support the idea that the needle did not puncture the dura, the procedure was performed under fluoroscopic guidance; no cerebrospinal fluid leakage was observed and a good epidurogram was obtained. Headache typically occurs within the first 48 h after dural puncture, with 90% occurring within the first three days. In the largest follow-up study of patients with postdural puncture headaches, nearly three-quarters of headaches resolved within seven days. In our case, headache became the main symptom more than seven days after the procedure and rather than resolving, it increased in intensity. Spontaneous intracranial hypotension is caused by spontaneous cerebrospinal fluid leaks that are generally undetected as no localised symptoms are present. In the majority of cases, the precise cause of these leaks remains unknown but an underlying weakness of the spinal meninges is suspected. However, a history of trivial trauma before the onset of symptoms can be elicited in about one third of patients. If these patients go on to surgical treatment, a variety of dural defects are observed, ranging from small holes to complex meningeal diverticula . Patients with spontaneous intracranial hypotension usually have a fronto-occipital orthostatic headache, although a wide variety of other headache patterns are reported. A number of other symptoms have been documented including tinnitus and balance problems. Treatment for spontaneous cerebrospinal fluid leak ranges from conservative management (bed rest, oral rehydration and an abdominal binder) to surgical closure of the defect. The mainstay of treatment is epidural blood patching, which is also a part of the diagnostic criteria published by the International Classification of Headache Disorders.

Recently, four patients have been reported with a syndrome of hypoactive-hypoalert behaviour associated with spontaneous cerebrospinal fluid leak. The patients in these reports are described as suffering from persistent headache, fatigue with somnolence, general apathy, impaired attention and stereotypical motor activity. The patients’ ages ranged from 40 to 60 years, and the duration of symptoms was up to four years. Notably, the patients presented only after the continued concerns of relatives. Neuro-psychological testing showed broadly normal cognitive function, with mild impairment in attention and memory demonstrated in two patients. Three of the four patients (one was lost to follow-up) were treated with an epidural blood patch, with almost immediate resolution of their symptoms. In each case, however, the effect was not permanent and the procedure was repeated after 5–8 months.

The symptoms initially experienced by our patient were in keeping with a diagnosis of postdural puncture headache. However, as time progressed the symptoms worsened rather than improved and the clinical picture became closer to that described in patients with spontaneous intracranial hypotension. One of our patient’s most concerning features was the somnolent hypoactivity that became worse as time progressed. Whilst not identical in description to the hypoactive-hypoalert behaviour described above, there were similarities, not least the dramatic resolution of symptoms with epidural blood patching. We do not think our patient’s symptoms can be explained as a ‘simple’ postdural puncture headache. We feel that the arguments relating to the implausibility of initial dural puncture by our intervention needle are too strong, and combined with the delay in presentation, we believe that there is insufficient evidence to support this diagnosis. Despite a lack of definitive radiological evidence, but given our patient’s remarkable response to epidural blood patching, her symptoms are more aligned with those described in spontaneous intracranial hypotension. We hypothesise that the aetiology of our patient’s clinical syndrome was as follows. We believe that the bolus of local anaesthetic and steroid mixture injected directly through the needle hub may have produced sufficient force to cause a small tear in the dura at a distal site, perhaps with a pre-existing anatomical weakness. It is conceivable that the shearing force produced by the hydrodissection could result in a tear and produce a clinical picture more in keeping with that of spontaneous intracranial hypotension. Our patient did not undergo myelography on first presentation, as this is not common practice in our hospital, but a successful epidural blood patch was performed, which is considered diagnostic. On second presentation, CT myelography was performed and was negative. However, it is thought that slow-flowing and intermittent leaks may evade detection due to either limitations in resolution or insufficient time between contrast administration and image acquisition.

Whilst the definitive cause of our patient’s condition remains unknown, treatment ith an epidural blood patch was effective. Given the same collection of symptoms, and despite an uncertain aetiology, we would recommend early intervention with an epidural blood patch. Publication History Article first published online: 31 JAN 2012
Accepted: 17 October 2011