LUMBAR puncture is usually contraindicated in patients with increased intracranial pressure or space-occupying lesions because lumbar cerebrospinal fluid (CSF) drainage can increase the pressure gradient between the supratentorial and infratentorial compartments and thus result in rapid herniation. 1,2Some authors have also suggested that epidural anesthesia is dangerous in patients with space-occupying lesions not only because of the risks associated with accidental dural puncture, but also because epidural drug and fluid injection can increase intracranial pressure. 3,4
We present a case of a pregnant woman with an occult brain tumor who received an epidural anesthetic for relief of labor pain, which was complicated by an inadvertent lumbar puncture. Several hours after delivery, she had fatal transtentorial brain herniation.
A healthy, 35-yr-old woman (primigravida) at 38 weeks’ gestation was admitted to the labor ward of a local obstetric hospital in spontaneous labor. The course of pregnancy had been uneventful. At the time of admission, she was afebrile, her pulse rate was 86 beats/min, and her blood pressure was 136/104 mmHg. Laboratory studies were normal.
Epidural analgesia was initiated during the first stage of labor. An initial attempt was made with an 18-gauge Tuohy needle at the L3–L4 interspace, with the patient sitting. However, an accidental dural puncture occurred; 3 ml air being used for “loss-of-resistance” testing was injected into the intrathecal space. She reported no symptoms. A second attempt was performed at the L2–L3 interspace, and an epidural catheter was successfully introduced into the epidural space. After an uneventful test dose, 12 ml bupivacaine, 0.25%, was injected over 20–30 s. This injection was accompanied by a transient headache. Analgesia was satisfactory. Two further injections with same volume and concentration of bupivacaine were administered during the first stage of labor.
During labor, mean arterial pressure varied from 140 to 160 mmHg. The process of labor was uneventful, except that the patient reported occasional severe headache during labor, particularly during the second stage. Estimated blood loss was approximately 150 ml. Overall, the first stage of labor lasted approximately 8 h, and the second stage lasted approximately 20 min. Delivery was uneventful.
After delivery, she was sent to the ward for routine observation. However, 2 h after delivery, a marked decrease in her level of consciousness was noted. Her pupils were equal and reacted normally to light, and there were no focal neurologic deficits. Blood pressure was 156/112 mmHg; heart rate was 102 beats/min. No evidence of postpartum hemorrhage was noted. Laboratory studies were normal. However, 1 h later, left pupillary dilatation was noted. Her trachea was emergently intubated, and she was transferred to our hospital, arriving in the emergency room approximately 4 h after delivery.
At the time of arrival, blood pressure was 74/35 mmHg, and heart rate was 108 beats/min. Neurologic examination revealed deep coma (Glasgow Coma Score = 3). Her pupils were fixed and dilated bilaterally. Brain stem reflexes were absent. Hemodynamic resuscitation was accomplished with intravenous saline and a dopamine infusion. Computed tomography revealed a slightly hyperdense tumor in the left hemisphere with marked perifocal edema (fig. 1). There was a left-to-right midline shift of forebrain structures; tentorial herniation was also noted. A small amount of intracranial air was noted.
Because of her poor neurologic and hemodynamic status, her family declined emergent craniotomy. She was moved to the neurosurgical intensive care unit for further care and died 10 days later. An autopsy was not performed.
The young woman described in this report had a tragic, fatal complication of her pregnancy. Although it is possible that her death may have been due solely to the combination of labor, delivery, and her tumor, we believe that her anesthetic may have contributed to this outcome. First, she had a large but unknown brain tumor. The only possible symptom attributable to this lesion was headache at the time of her initial epidural injection, and recurrent headaches during labor. Her hypertension may have been in response to increased intracranial pressure or, even if unrelated, may have lead to worsened intracranial hypertension and cerebral edema. Episodic increases in venous pressure, due to Valsalva maneuvers, could have been another contributing factor. The CSF leak induced by the inadvertent lumbar puncture then resulted in an increased transtentorial pressure gradient, which resulted in herniation.
The role of the epidural drug injection in her outcome is less clear. Epidural fluid injection can increase intracranial pressure, probably because of compression of the lumbar thecal sac with translocation of CSF into the intracranial space. 3,4However, in this case, such changes might have actually delayed the development of herniation. The increased epidural pressure produced by the injection may have slowed the loss of CSF. By increasing infratentorial pressure, the pressure gradient between the supratentorial and infratentorial space may also have been reduced. The 11-h interval between her accidental lumbar puncture and deterioration may have been the result of these temporarily “beneficial” changes. No additional epidural injections were performed after delivery. As fluid in the epidural space was absorbed, her CSF leak may have increased, leading to a worsened transtentorial pressure gradient and subsequent herniation.
How can we prevent this devastating complication? Certainly, epidural anesthesia would not be considered in patients with known mass lesions because of the risk of unintentional dural puncture. The difficulty is identifying patients with occult lesions. It is possible that a more carefully obtained history and physical examination might have been helpful. If there is an unexplained history of worsening headache, intractable vomiting, seizure, focal neurologic signs, altered consciousness, or papilledema, computed tomography should be performed. In this case, no symptoms were evident (other than headache). The delayed recognition of her changing neurologic status—and hence a delay in the institution of aggressive treatment (intubation, hyperventilation, osmotic diuretics, and others)—almost certainly contributed to the poor outcome. Even if preanesthetic diagnosis is not possible, closer postdelivery observation may be lifesaving.