Key Factors Determine Success of Epidural Patch for CSF Leaks

Pauline Anderson

May 02, 2018

VANCOUVER, British Columbia — Researchers are gaining more insight into the factors that best predict the efficacy of an epidural blood patch in patients with spontaneous cerebrospinal fluid (CSF) hypovolemia.

New research shows that volume of blood injected, number of injection sites, and site-directed strategies  were significantly correlated with patch efficacy, defined as lasting at least 3 months.

Dr Gabriel Pagani-Estevez

"We also showed that there are certain brain angles on the MRI scan that will help you know in advance if a patient is likely to have a stormy course or is more likely to do okay," lead author, Gabriel Pagani-Estevez, MD, a pain fellow at Mayo Clinic, Rochester, Minnesota, told Medscape Medical News.

The study was presented here at the American Academy of Pain Medicine (AAPM) 2018 Annual Meeting.

"Brain Sag"

Spontaneous CSF hypovolemia — or a CSF leak — is uncommon but not rare, said Pagani-Estevez. Patients with this condition lose fluid around the brain stem and can develop "brain sag," where the brain "is sinking." A classic symptom is orthostatic headache.

Spontaneous CSF hypovolemia is often misdiagnosed or not promptly diagnosed. Patients can go for years without a proper diagnosis, he said.

This may be because over time the orthostatic nature of the headache can be lost, and it can just be a chronic headache.

Also, patients can present with varying symptoms. "They can have cranial nerve palsies, they can have a lot of hearing manifestations, they can have neck pain without a headache."

Epidural blood patches are surgical procedures that use autologous blood to close one or more holes in the dura mater of the spinal cord.  The injections increase pressure in both the epidural space and in the CSF space and improve brain sag.

Although these patches are a safe and effective treatment for spontaneous CSF hypovolemia, clinical and procedural variables predictive of their efficacy have not been rigorously described.

In this retrospective case-series study, researchers identified 202 patients who had received a total of 604 epidural blood patches. Most patients needed more than one blood patch; about half had at least two, 47% had three, and 20% had five.

"So for most patients, it's not one and done," said Pagani-Estevez.

This suggests that these patients "tend to be kind of refractory" and for them, the patches aren't always curative. "This may be a chronic illness and people may need constant epidural blood patches."

Alternatively, he said, "there could be a cumulative effect whereby a number of patches are needed before a longer-lasting remission of symptoms is seen."

About 80% of the patches in the study were blood only, and the remainder involved some form of fibrin.

As for injection strategy, 78% of patches were single-level patches (involving a single needle), 12% were bilevel (involving two different needles and levels), and 9% were multilevel (three or more levels). Of the single-level patches, 58% were in the lumbar spine.

The average period of relief — defined as the time between when the patch started to work and when it started to wear off — for the entire study group was only 2 weeks, said Pagani-Estevez.

Univariate and multivariate statistical analyses showed that higher blood volume predicted patch effectiveness (odds ratio [OR], 1.64; P < .0001).

"So the more volume you put in, the more likely the patch is to be effective," said Pagani-Estevez.

Adding more than a single-level patch increased the chances of success. For example,  the OR was 3.17 (P < .0001) for bilevel injections and jumped to 117 (P < .0001) for the multilevel patch.

"We showed that when you add more levels to the procedure itself, the patch is more likely to be effective and longer lasting," said Pagani-Estevez. "So when you're doing the procedure, if you add multiple needles and inject along multiple sites of the epidural space," the outcome appears to be better.

This is likely explained "on the basis of a volume effect alone, since the volume is the only variable that retained its significance on multivariable model," added Pagani-Estevez.

MRI findings of brain sag (eg, midbrain-pons angle < 47° and vein of Galen/straight sinus angle < 58°) and four or more MRI abnormalities were negative predictors of the patch working at 3 months.

Best Route of Administration

The study results also suggest that with blood-only injections ؙ—  those without fibrin ؙ— there is less likelihood of a patch working at 3 months. In addition, the results imply that a transforaminal route of administration is more likely to lead to an effective patch.

"But it should be stressed that the patches incorporating fibrin, or patches incorporating the transforaminal route of administration, were overwhelmingly more likely to be multilevel and high-volume patches, so there's a bit of a confounding effect when interpreting those variables," commented Pagani-Estevez.

Site-directed patches were more effective than nontargeted patches (OR, 8.35; P = .033).

"We were able to show that if you know where the leak is, and you deliver the injectate right at the site of the leak, it's a lot more likely to be effective," said Pagani-Estevez.

If the location of the leak is unknown, Pagani-Estevez said he would "advocate" for targeting the low thoracic and lumbar spine.

One reason for this is that most leaks are in the thoracic spine, particularly lower down. Also, a greater volume can be accommodated lower down in the spinal canal, "and we showed that volume is a significant predictor of patch efficacy," said Pagani-Estevez.

Rebound intracranial hypertension is another possible predictor of patch responsiveness. In this review, 14 patients developed this condition.

"The patches that caused it were associated with about 140 days of relief," said Pagani-Estevez.

"If a patient develops these higher-pressure symptoms after the patch, that's a harbinger of a successful patch and not of a patch failure necessarily; if you keep on patching them, it will make the problem worse."

"Exciting Advance"

Commenting on the findings for Medscape Medical News, Patrick J. Tighe, MD, associate professor of anesthesiology, and program director, Perioperative Analytics Group, University of Florida, Gainesville, said the study addresses an increasingly recognized phenomenon.

While more attention is being paid to spontaneous CSF hypovolemia, there has been "a relative lack of evidence about who benefits best from which treatment," said Tighe.

"Epidural blood patches are not risk free, and so we'd like to show that patients are getting better and we are minimizing risk and unnecessary care."

Possible risks of these patches include bleeding and infection, and patients receiving a blood thinner may develop an epidural hematoma, which, although exceedingly rare, can cause permanent nerve injury.

"This is an exciting advance" in that the authors have collected data on factors affecting the effectiveness of epidural blood patches in patients with spontaneous CSF hypovolemia, said Tighe. "They have done a great job in beginning to address a pretty complicated issue."

No funding sources or conflicts of interest were reported.

American Academy of Pain Medicine (AAPM) 2018 Annual Meeting. Abstract 162. Presented April 27, 2018.

For more Medscape Neurology news, join us on Facebook and Twitter


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.