OJAI, CA—Spontaneous CSF leaks are treatable, often misdiagnosed, and can cause a neurologic syndrome that may include headache, nausea, and tinnitus. Spinal fluid leaks also can lead to serious complications, including seizures. Patients may have a CSF leak for years or decades before it is diagnosed.
Although CSF leaks may not be readily apparent on imaging, a suspected CSF leak is important to consider because it is fixable, said Dr. Carroll, Assistant Professor of Anesthesiology and Perioperative and Pain Medicine at Stanford University Medical Center in California and a member of the Stanford CSF leak program.
Postdural Puncture Headache Versus Spontaneous CSF Leaks
A spontaneous CSF leak and the clinical syndrome that it causes may be confused with a postdural puncture headache.
With a postdural puncture headache, a patient usually has a single leak site in the dorsal dura. “There is up to a 90% response to a single epidural blood patch. Its natural history is generally well understood and benign. It is rarely mysterious, and it is ultimately fixable,” said Dr. Carroll. In contrast, “a spontaneous CSF leak is often mysterious in terms of the onset, cause, and diagnosis.”
The natural history of CSF leaks is poorly understood. The percentage of patients whose spontaneous leaks seal on their own or whose leaks cause a catastrophe (eg, coma, seizures, or hematomas) is not known. Between 30% and 40% of patients with spontaneous leaks have leaks from multiple sites at diagnosis. “With a spontaneous leak a dural rent is more likely in the ventral dura, anterior to the spinal cord or at the nerve root, making the dura less amenable to patching. A single patch at the correct spinal level will fix the problem only 30% of the time. With multiple patches, the success rate can approach 65% to 75%.” If the first epidural blood patch fails, it should be repeated. Directed epidural blood patches, placement of fibrin sealant, and surgical treatment are other treatment options.
Most patients with post-puncture CSF leaks have classic orthostatic headaches. The orthostatic headaches from spontaneous leaks are often atypical, however, in that patients may not feel better immediately when they lie down and worse when they stand, Dr. Carroll said. Headaches may occur late in the day after prolonged upright time or with exertion. In addition, patients may “go from having an orthostatic headache to having a terrible headache all the time, regardless of what position they are in.”
Nausea and Vomiting
Nausea and vomiting can be the main symptom of a CSF leak. Dr. Carroll described a patient with complex regional pain syndrome who underwent a spinal cord stimulation trial. Afterward, she had a postdural puncture headache and received an epidural blood patch. “After that, she developed vomiting up to nine times a day.” A CSF leak was not visible on the first CT myelogram but was apparent on the second. The leak was “so small it was missed by the slice thickness” of the first CT myelogram, said Dr. Carroll. She ultimately required surgery to fix the leak. The patient improved, although she continued vomiting three times per day.
CSF leaks may cause tinnitus. “You can get ringing in the ears when you have migraine,” Dr. Carroll said. But if patients have tinnitus when they are not having headaches, “you should be thinking that there is something else going on.” Data suggest that CSF fluid is connected to inner ear fluid so a change of pressure in CSF changes inner ear pressure, and patients with high or low CSF pressure may get tinnitus.
Other symptoms may include neck pain and fatigue. “I have had the parents of patients tell me that the most remarkable thing that they see when we patch their sons or daughters is how they are bouncing around the house,” he said. Many patients complain of difficulty with concentrating, task persistence, and other nondescript, nonfocal neurologic symptoms.
Imaging of patients with CSF leaks often initially is read as normal, and MRI is not an adequate evaluation in the high clinical suspicion of a leak, Dr. Carroll said. “It is a good place to start, because if you see a leak on your MRI, maybe you do not have to get a CT myelogram,” he said. “But if you have a clinical suspicion of a leak … you should pursue that in the face of your radiologist telling you that there is nothing.”
Schievink et al in 2007 looked at several years of data from an emergency department to assess how often imaging findings consistent with CSF leaks were missed. They reviewed MRIs of patients with headache to look for evidence of intracranial hypotension, and then compared the number of CSF leaks with the number of subarachnoid hemorrhages seen during the same time. They found that for every subarachnoid hemorrhage, there were approximately 0.5 CSF leaks (23 subarachnoid hemorrhages and 11 CSF leaks). The results suggested that spontaneous intracranial hypotension is more common than previously thought and its diagnosis in emergency departments is problematic. The 11 people with MRI evidence of intracranial hypotension subsequently were diagnosed with CSF leaks and treated. None were diagnosed at the time of the MRI while in the emergency department.
Causes of Leaks
The four main causes of CSF leaks are medical procedures; whiplash; bony, sharp calcifications penetrating the dura; and genetic disorders of connective tissue.
Webb et al conducted a study to evaluate headaches in patients who had a known wet tap (ie, unintentional dural puncture) after a labor epidural. The researchers reviewed quality assurance data in an obstetrics anesthesia division and identified 40 patients who had known wet taps and 40 controls who had received an epidural without a wet tap during the same week and were matched for age and weight. Investigators contacted patients between 12 and 24 months later (mean, 18 months) and asked them about the incidence of chronic headache. The incidence of chronic headache in controls was 5% versus nearly 30% in patients who had had a wet tap. The investigators then compared patients who were managed conservatively (ie, they did not receive an epidural blood patch) versus patients who were managed with a blood patch. “If you got a blood patch, your risk of having a chronic headache 18 months later was only half as much as if you did not get a blood patch,” he said.
Connective Tissue Disorders and Calcifications
Because connective tissue disorders are associated with CSF leaks, headache physicians should determine patients’ Beighton Hypermobility Scores, Dr. Carroll said. The score is derived from a simple test that assesses joint hypermobility. For instance, patients receive a point if they can touch their thumb to their wrist or straighten their elbow more than 10° beyond 180°. A score between 3 and 5 raises suspicion that the patient might have a hereditary disorder of connective tissue. Cataracts at an early age, being unusually tall or short, degenerative disc disease, and history of aneurysm also are associated with an increased risk of CSF leaks.
With regard to calcifications, Dr. Carroll described a patient whose main complaint was confusion upon standing too long. The patient also had neck pain. They determined that he had a calcified bone spur that was puncturing the spinal cord, causing a leak.
Trauma and whiplash can cause leaks. Researchers in Japan studied 66 patients with chronic whiplash-associated disorders (ie, they had a whiplash accident and complained of neck and head pain, as well as difficulty with fatigue or memory). The patients were given a radionuclide cisternogram. Thirty-seven of the 66 patients had imaging that was positive for a CSF leak. “After being patched, roughly half the people who were found to have a leak went back to work, whereas they had not been working before,” he said. Another study found that 10% of people with brachial plexus injuries have spinal fluid leaks.
Overlap With POTS
The fact that Ehlers-Danlos also is associated with postural orthostatic tachycardia syndrome (POTS) raises the possibility that patients with CSF leaks may be misdiagnosed as having POTS.
“Why should a hereditary disorder of connective tissue be associated with the only two known conditions that are associated with feeling worse when you are up?” Dr. Carroll asked. Among patients with POTS, 60% have headaches, and many have dizziness and nausea. Dr. Carroll asked the POTS clinics at Stanford to refer patients with POTS, headache, and Ehlers-Danlos syndrome to him. The first referred patient’s history was consistent with a CSF leak. She had been passing out and had severe headaches for more than 10 years. Although her initial imaging was read as negative for CSF leaks, and an MRI showed no signs of intracranial hypotension, “when we patched her, in fact, she got better.” Subsequently, more patients diagnosed with POTS have been referred to the CSF leak program.
Patients initially may be diagnosed as having chronic fatigue syndrome, fibromyalgia, or irritable bowel syndrome when a CSF leak is causing their symptoms. It is a tragedy when patients “have a fixable leak and … nothing is done to treat that underlying problem,” Dr. Carroll said.
Ishikawa S, Yokoyama M, Mizobuchi S, et al. Epidural blood patch therapy for chronic whiplash-associated disorder. Anesth Analg. 2007;105(3):809-814.
Schievink WI. Spontaneous spinal cerebrospinal fluid leaks and intracranial hypotension. JAMA. 2006;295(19):2286-2296.
Schievink WI, Maya MM, Moser F, et al. Frequency of spontaneous intracranial hypotension in the emergency department. J Headache Pain. 2007;8(6):325-328.
Webb CA, Weyker PD, Zhang L, et al. Unintentional dural puncture with a Tuohy needle increases risk of chronic headache. Anesth Analg. 2012;115(1):124-132.