EPIDURAL blood patch (EBP) is now recommended as the treatment of choice for managing postdural puncture headache (PDPH). 1–3Choi 2noted an initial success rate of 91% and a long-term success rate of 61% in managing PDPH with EBP. In a prospective study of 504 patients undergoing EBP for PDPH, however, Safa-Tisseront et al . 3observed less initial success in managing PDPH with EBP and reported complete relief of symptoms in 75% of patients, incomplete relief in 18%, and failure in 7%. The currently recommended volume of autologous blood for EBP is 10–20 ml 1,2; however, repeated EBPs for refractory PDPH have used larger volumes of autologous blood. 4,5We report a case of permanent spastic paraparesis with cauda equina syndrome after EBP with 30 ml of autologous blood for the successful management of PDPH.
A 52-yr-old woman with a past history of hypothyroidism, L5-S1 discectomy in 1986, and prolonged cervical pain and facial numbness from an automobile accident in 1998, underwent outpatient lumbar puncture (LP) in 1999 to rule out multiple sclerosis (table 1). The LP was performed atraumatically with a 22-gauge, noncutting spinal needle, at the L2–L3 interspace and the upper pole of the earlier laminectomy scar. Cerebrospinal fluid (CSF) analysis, clotting studies, serum chemistries, and hematologic indices were within normal limits. Postural, bilateral, frontoparietal headache with occipital radiation, photophobia, and nausea began the morning after the LP, and worsened over the next 2 days. On the third day after LP, the patient, who remained afebrile, underwent outpatient EBP for PDPH, performed by an anesthesiologist in the emergency department. The EBP was performed atraumatically in the left lateral decubitus position with a 17-gauge Tuohy needle inserted in the midline at the L2–L3 previous LP site through a 1% lidocaine skin wheal. A loss-of-resistance technique with preservative-free normal saline and no air was used to detect epidural space entry. On epidural space entry, no paresthesias, blood, or CSF were noted. No air or local anesthetics were administered into the epidural space. During slow epidural injection of 30 ml of autologous blood, the patient reported no low back, buttock, or lateralizing leg pain. The patient was then placed at supine bedrest, reported 90–95% improvement in her postural headache and associated symptoms, and was discharged from the emergency department 30 min after the EBP.
Although her postural headache improved, the patient experienced low back pain after returning home. Over the next 48 h (days 1–2), the low back pain worsened and was associated with lumbar and left leg muscle spasms and radiating radicular pain in the left leg and foot. On the third day after EBP, the patient was believed to have routine, resolving sequelae from successful EBP on examination by her personal physician, who detected no neurologic signs and prescribed oral gabapentin for pain. On the fourth day after EBP, the patient was again believed to exhibit sequelae of successful EBP on examination by a consultant anesthesiologist, who did not perform the EBP. The patient continued to experience low back and left leg radicular pain, which limited ambulation. The patient continued taking gabapentin for pain and avoided anticoagulants, nonsteroidal antiinflammatory drugs, and herbal supplements. On post-EBP day 12, urinary incontinence and paraparesis of the left leg developed in the patient. An immediate magnetic resonance imaging (MRI) study of the lumbar spine showed a sausage-shaped mass in the posterior lumbar spinal canal, extending from L2–L4, and causing extrinsic spinal cord compression at L2–L3. Despite improvement in back pain and leg weakness after 24 h of intravenous dexamethosone, as recommended by the neurosurgical consultant, the patient continued to have back pain and left foot weakness. On post-EBP day 13, the neurosurgeon recommended lumbar laminectomy with evacuation of lumbar epidural hematoma.
Extending the earlier lumbar laminectomy incision, the neurosurgeon performed a laminectomy from L2–L4 and observed an organized subdural hematoma, sandwiched between the dura and arachnoid mater. Although the MRI had suggested a posterior epidural hematoma, laminectomy uncovered a subdural hematoma that extended from L2–L4, and compressed the arachnoid mater against the exiting spinal nerves bilaterally. The subdural hematoma was evacuated, and the dura and other soft tissues were closed in layers. Postoperatively, the low back pain and muscle spasms improved, but severe left hip and leg pain continued. A spinocutaneous fistula with a CSF leak developed during ambulation on the fourth postoperative day, and was managed medically with antibiotics and fluid restriction. Lumbar MRI on the seventh postoperative day revealed a cystic CSF collection from L2–L5. The recent laminectomy incision was reexplored by another neurosurgical consultant; several dural leaks were identified, oversewn, and resealed with fibrin glue.
Over the next year, bilateral leg weakness, opisthotonic muscular spasms in the lower back and left leg, urinary incontinence, and sexual dysfunction, consistent with spastic paraparesis and cauda equina syndrome developed in the patient. The muscular spasms were associated with painful accentuated lumbar lordosis and spastic inversion of the left foot. The diagnosis of multiple sclerosis was excluded from the current illness by normal CSF analysis, absence of cerebral and spinal sclerotic plaques on multiple neuraxial MRIs, and absence of ophthalmoplegia or other cranial nerve findings.
On referral to our pain clinic in November 2000, the patient was taking several oral medications for anxiety and pain (citalopram, alprazolam, propoxyphene), and for painful spasticity (diazepam, baclofen, tizanidine). A physical examination revealed a wobbly, cane-supported gait with left foot drop, bilateral sacroiliac tenderness, well-healed lumbar laminectomy scars, patellar hyporeflexia, and Achilles areflexia. The patient began taking a new combination of oral medications for anxiety and chronic pain (zolpidem, tramadol, gabapentin, rofecoxib) and for painful myospasticity (baclofen, tizanidine), and was gradually weaned from the propoxyphene and benzodiazepines. With the addition of aggressive physical therapy and plantar left foot support, the patient's pain and mobility improved over the next 6 months with fewer myospastic attacks.
Epidural blood patch (EBP) is the currently accepted treatment of choice for postdural puncture headache (PDPH) because of its high initial success rates (over 70%), low failure rates (less than 10%), and infrequent complications. 1–3The ideal volume of blood to be injected during EBP is between 10 and 20 ml, although some investigators have recommended larger volumes as directed by the onset of backache, neckache, or radicular pain. 1,3–5Bart and Wheeler 6compared treatment with EBP using 10 ml of autologous blood with epidural saline patch (ESP) of 30 ml in patients with PDPH after spinal anesthetics with 25-gauge needles and epidural anesthetics with 17-gauge needles. In the PDPH populations compared, EBP was significantly superior to ESP in resolving PDPH. 6Taivainen et al. 5compared EBP for PDPH using 10 ml of autologous blood to EBP for PDPH with 11 to 15 ml of blood and found no difference between the 2 groups.
The most common complication associated with EBP is low back pain. 7Other rare, transient complications have included aseptic meningitis, 8lumbovertebral syndrome, 9radicular pain, 10bradycardia, 11fever, 12and seizures. 13Spinal hematoma is a known complication of dural puncture. It may occur spontaneously, and can cause transient or permanent paraparesis. 14,15Spinal hematoma has also occurred after dural puncture in anticoagulated patients, 15,16and in patients with renal failure, 17liver failure, 18and thrombocytopenia. 18Tekkok et al. 19reported the surgical evacuation of a spinal epidural hematoma in a woman without coagulopathy, who underwent a series of six EBPs to manage PDPH after repeated epidural phenol injections for chronic pain.
In this case, EBP was successful in relieving PDPH, was associated with common side effects, and was not associated with excessive bleeding, paresthesias, or recognized dural puncture. The delayed presentation of severe neurologic findings in this case was inconsistent with pioneering animal investigations conducted by DiGiovanni et al . 20that demonstrated disappearance within days of 2 ml of autologous blood injected epidurally in goats. The 30-ml EBP clotted, initially relieving PDPH; defibrinated over time; and reorganized into a dense mass, possibly more seroma than hematoma, by post-EBP day 12. In addition, the contribution of leaking CSF to the mass effect may have expanded the remaining elements of the EBP and the phagocytic process, contributing progressively to the compression of the terminal spinal cord and cauda equina. The initial MRI study demonstrated a lumbar mass effect that caused extrinsic cord compression from L2–L4, consistent with a posterior spinal epidural hematoma. However, spinal decompression showed the hematoma to be sandwiched between the dura and arachnoid mater, tamponading any possible CSF leakage. Aldrete and Brown 21also reported subdural hematoma after EBP without observed dural puncture, but the EBP was injected through an epidural catheter presumed to have migrated intrathecally. Spinal reexploration for CSF fistula revealed multiple CSF leaks at previous LP and laminectomy sites. These dural rents may have opened subdural tunnels for the EBP to follow. Postlaminectomy fibrosis with obliteration of the epidural space may also have compromised epidural space identification, predisposing to dural puncture, as occurred in the case of subdural hematoma after repeated epidural phenol injections with presumed fibrosis. 19
In conclusion, we have reported a case of permanent spastic paraparesis with the cauda equina syndrome after EBP for PDPH. The only steps in the EBP process in this case that were unconventional included an EBP volume of 30 ml and a short post-EBP supine observation period of 30 min, instead of the recommended 1–2 h. 5These steps were not, however, inconsistent with the experiences of other investigators, managing similar patients with PDPH by EBP. 5,6The low back and radicular pain reported by the patient during the week after EBP were assessed by two physical examinations and considered consistent with typical post-EBP sequelae. The dramatic onset of left leg weakness and urinary incontinence 12 days after EBP precipitated immediate radiographic and neurosurgical evaluation for spinal compression.
Epidural blood patch for PDPH should be limited to 10–20 ml of autologous blood, injected as close to the initial dural puncture site as possible, with injection halted by patient reports of significant low back or leg pain. Epidural blood patch should be followed by a 1 to 2 h period of supine bedrest to allow the EBP to clot and adhere locally. 5Anesthesiologists must remain vigilant for spinal compression after EBP for PDPH, especially in postoperative laminectomy and postphenol neurolysis patients at greater risk of dural puncture. 19Surgical decompression of the terminal cord and cauda equina must be immediate, and complete recovery or transient neurologic outcomes cannot be assured.