Intracranial Subdural Hematomas and Cerebral Herniation after Labor Epidural with No Evidence of Dural Puncture

The patient was a previously healthy, primiparous 40-yr-old Chinese woman pregnant with twins. Her medical and surgical history was unremarkable, with the exception of a history of hepatitis B. She took no medications and reported an allergy to aspirin leading to rash. The patient presented to the obstetric service at 38 weeks’ gestation for scheduled induction of labor and subsequently requested epidural analgesia. During the epidural placement, the patient was in the seated position, was prepared and draped in the usual sterile fashion, and received a 1% lidocaine skin wheal. The L4–L5 epidural space was then accessed using a 17-gauge Tuohy-Weiss needle using a loss of resistance to saline technique. The epidural placement was uncomplicated: It occurred on the first attempt without paresthesia and did not result in a visible cerebrospinal fluid (CSF) leak or any other clinical signs of dural puncture. Adequate epidural analgesia was easily attained with the usual dosages of local anesthetic.

Despite 2.5 h of maternal pushing, the presenting twin remained in the occiput-posterior position at 2+ station, and the decision was made to proceed with cesarean delivery. The delivery of her twins was subsequently uncomplicated, and the patient was discharged on postoperative day 3 with visiting nursing care. At the time of discharge, the patient was without specific complaints.

Three weeks later, the patient presented to the labor floor with a vague history of worsening headache, not clearly positional, which she said had been developing since her discharge. She reported a near-syncopal episode and the presence of blurry vision, but she denied weakness, numbness, nausea, or vomiting. Upon arrival to the hospital, the patient had a witnessed seizure in the elevator. Urgent, noncontrast computerized tomography of the head revealed large bilateral subdural hematomas containing acute and subacute blood (fig. 1). There was evidence of subfalcine and uncal herniation, entrapment of the right lateral ventricle, and mass effect with 7 mm of shift of midline structures to the right (fig. 1).

A neurosurgical consultation was immediately obtained, and urgent burr-hole craniotomies were performed for evacuation of the subdural hematomas. Other than a mild headache, the patient felt well and was without neurologic deficit in the postoperative setting. Despite the absence of a documented dural puncture or the presence of ongoing symptoms, the treating neurosurgeon requested that an epidural blood patch be performed to prevent further CSF loss and reaccumulation of the subdural hematomas.

The day after evacuation, the patient was sent to our Pain Clinic for fluoroscopic-guided epidural blood patch. Despite the additional precaution of visualization under fluoroscopy, inadvertent dural puncture occurred at L3–L4 with a 20-gauge needle during the initial attempt, which was accompanied by the expected event of CSF fluid output under pressure. A second attempt at L4–L5 resulted in a successful epidural blood patch. The patient remained supine, with the head of the bed flat for 2 days. Subsequent neuroimaging revealed no significant interval change or evidence of new intracranial hemorrhage, subdural hemorrhage, or subdural fluid collection. The patient was subsequently discharged on postoperative day 5 without deficit. One month later, computerized tomography revealed no recurrence of hematoma or mass effect (fig. 2).

The differential diagnosis of headache in the peripartum population includes muscle tension headache, migraine, preeclampsia or eclampsia, subarachnoid hemorrhage, stroke, tumor, and cerebral venous thrombosis.12,13Post–dural puncture headache and meningitis must also be considered in the setting of recent neuraxial analgesia or anesthesia. Intracranial subdural hematoma is a rare but well-documented complication of epidural anesthesia.1–5In these published reports, dural puncture was clearly noted at the time of epidural placement. In the current report, however, the patient had an apparently uncomplicated epidural placement, reported her headache 3 weeks postpartum, and had a nearly fatal neurologic outcome.

Given the absence of a frank wet tap in this patient, it was important to consider other possible etiologies of her bilateral subdural hematomas, such as traumatic brain injury, coagulopathy, intense Valsalva, or preexisting intracranial hypotension. The absence of a history of domestic abuse or external bruising, as well as the bilaterality of the hematomas, argues against traumatic etiology. An unrecognized coagulopathy that increased her risk of bleeding was also considered. If that had been the case, however, she would likely have been at risk for excessive bleeding at the time of her cesarean delivery and for development of a spinal epidural hematoma, neither of which occurred.

The question of whether the intense Valsalva during the 2.5 h of pushing could itself explain the subdural bleeding was raised. Spontaneous subdural hematoma is typically reported in elderly patients presumably due to brain atrophy and stretching of dural veins. It may also occur in younger individuals with preexisting spontaneous intracranial hypotension.14In this patient population, there may be no detectable CSF pressure during lumbar puncture with a 20- or 22-gauge needle. In the event of dural compromise with the 17-gauge needle used during epidural placement in such a patient, it is unclear whether the typical signs of a wet tap would be evident.

The intracranial pressure in this patient underwent a dynamic course. Intracranial hypotension was likely a precipitating cause of the subdural hematoma formation, either from preexisting disease or induced by occult dural tear. Accumulation of subdural blood and potential sealing of the CSF leak led to intracranial hypertension, resulting in herniation and ultimately relieved by craniotomy. At the time of epidural blood patch placement, however, the patient did demonstrate sufficient CSF pressure to manifest an obvious wet tap. This fact, in addition to her lack of symptoms before labor and epidural placement, argues against an underlying spontaneous intracranial hypotension. The most likely explanation is that this patient had an occult dural puncture that extended in the postpartum period, leading to the complication of rapid and significant CSF leak. This may have subsequently caused intracranial hypotension, stretching of the dural veins and subdural bleeding. The stoic nature of this particular patient, perhaps combined with a cultural unwillingness to complain, may have contributed to her late presentation.

The role of the epidural blood patch in this case is controversial. As first described by Gormley15in 1960, the blood patch was proposed to treat a postdural puncture headache by tamponading the CSF leak. It followed, then, that a blood patch could be used to prevent further CSF leak and accumulation or reaccumulation of subdural hematomas. This outcome is not always the case: Subdural hematomas have been found to develop after inadvertent dural puncture even when an epidural blood patch was placed.3Furthermore, blood patches themselves are not without known risks and have been reported to exacerbate neurologic symptoms of inadvertent dural punctures.16It has also been suggested that administration of an epidural blood patch induced seizures in a patient with a postpartum headache and an undiagnosed intracranial subdural hematoma.17 

Despite the use of fluoroscopy, our patient experienced the complication of inadvertent dural puncture during an attempted blood patch. One can only speculate as to why this documented dural puncture did not result in additional subdural bleeding. Possible explanations include the fact that the patient remained strictly supine for 2 days after the procedure (perhaps decreasing the likelihood of extending the dural tear and minimizing the gravitational forces on the potential CSF leak) and that a smaller needle was used for the blood patch than for the initial epidural (20 gauge vs.  17 gauge).

Postpartum headache can be relatively benign or alternatively can be a sign of significant pathology and should be evaluated promptly. Furthermore, in cases where the clinical history, signs, and symptoms are not clearly consistent with a dural puncture, neuroimaging should be strongly considered before epidural blood patch is performed. Detection of CSF leakage is also a diagnostic strategy that could be performed before further intervention.18In conclusion, this case exemplifies that serious neurologic sequelae such as intracranial subdural hematomas and cerebral herniation can occur even in the absence of obvious dural puncture.