Epidural Anesthesia in a Parturient with a Lumboperitoneal Shunt 

A 26-yr-old women, gravida 7, paragravida 1, was admitted at 32 weeks' gestation with severe preeclampsia. Her history included pseudotumor cerebri successfully treated 5 yr before with placement of a lumboperitoneal shunt at the L3-4 interspace. The preanesthetic evaluation revealed that she was obese (120 kg; body mass index, 41.5 kg/m²) and had an adequate airway anatomy and a normal platelet count (201 x 10^9/l).

Five days later, because of worsening clinical status, labor was induced with oxytocin. Repeated examination by the same anesthesiologist revealed significant changes in her airway since the initial evaluation. Her tongue obstructed the view of the soft palate and uvula, which had been easily visualized at the first examination. She also had marked facial edema. Her platelet count remained normal (290 x 10^9/l). The neurosurgeon who had placed the lumboperitoneal shunt had been consulted and believed that neuraxial anesthesia, including epidural catheter placement, was not contraindicated.

An open-ended, single-orifice epidural catheter was easily situated below the scar with a midline approach at the L4-5 interspace. Results of a test dose of 3 ml lidocaine. 1.5% with 5 [micro sign]g/ml epinephrine was negative. An additional 10-ml dose of bupivacaine, 0.125%, did not produce a detectable block. To determine the proper functioning of the catheter, 10 ml lidocaine, 1.5%, with 5 [micro sign]g/ml epinephrine was administered. This resulted in a dense T8-S1 sensory block on the left, T10-S1 on the right, and significant bilateral lower extremity weakness. Because she was not yet experiencing labor pain, the block was allowed to dissipate. She requested labor analgesia 3 h later. Injection of 8 ml bupivacaine, 0.25%, resulted in complete analgesia and a sensory block of T8-S1 on the left and T10-S1 on the right. A continuous of 10 ml/h bupivacaine, 0.125%, with 2 [micro sign]g/ml fentanyl was initiated. She remained comfortable for 7 h of adequate contractions, after which the obstetrician elected to perform cesarean delivery because of a nonreassuring fetal heart rate tracing and unchanged cervical dilation of 1 cm.

Epidural administration of 20 ml lidocaine, 2%, with 5 [micro sign]g/ml epinephrine, in divided doses, produced an inadequate right-sided block. The catheter was easily removed and replaced with a midline approach, this time above the scar at approximately T12-L1. The catheter was injected with 3 ml lidocaine, 1.5, with 5 [micro sign]g/ml epinephrine, and then 15 ml lidocaine, 2%, with 5 [micro sign]g/ml epinephrine, in divided doses, and 100 [micro sign]g fentanyl. This produced a dense bilateral sensory block to T8 before surgical drapes were placed. Although sacral nerve distribution was not tested, the patient was comfortable throughout surgery (90 min) and required no supplemental analgesia. A 1,535-g infant was delivered at 32 and 5/6 weeks. Apgar scores were 6, 8, and 9 at 1, 5, and 10 min, respectively. The epidural catheter was removed easily, and postoperative analgesia was provided via intravenous patient-controlled analgesia using morphine.

The mother's clinical status improved rapidly after delivery of the infant. But on the second postoperative day a fever of 40.8[degree sign]C developed. The obstetrician added clindamycin to a continuing regimen of ampicillin and gentamicin, both of which were initiated for a low-grade fever of unknown origin when labor was induced. The patient's antibiotics were changed to piperacillin-tazobactam when blood cultures grew Proteus mirabilis. Results of all other cultures and evaluations were negative. Abdominal and pelvic computed tomography ruled out abscess and revealed a normal lumboperitoneal shunt. The source of infection was believed to be genitourinary or gastrointestinal. The patient was discharged on postoperative day 8, was afebrile during oral antibiotic treatment, and had no further complications.

This case illustrates successful epidural anesthesia for labor pain management and subsequent cesarean delivery in a parturient patient with lumboperitoneal shunt for pseudotumor cerebri. Parturient patients with medically treated pseudotumor cerebri (i.e., those without a lumboperitoneal shunt) have successfully received various anesthetics, including epidural and spiral anesthesia for vaginal and operative delivery. [3,7]Some authors have described contraindications to neuraxial blockade in patients with lumboperitoneal shunts, [3]based largely on theoretical considerations. Concerns include trauma to the shunt and potential loss of local anesthetic into the peritoneal cavity (spinal anesthesia). They conclude that general anesthesia is preferable for cesarean section delivery of infants in patients with lumboperitoneal shunts.

No cases of trauma or disruption of the spinal portion of a lumboperitoneal shunt have been reported. Abouleish et al. [3]suggest that radiologic studies should be performed in the event of neuraxial anesthesia in patients with lumboperitoneal shunts. We thought this was unnecessary. Lumboperitoneal shunts are inserted into the dural sac at a low vertebral interspace. The tubing runs laterally, tunneled underneath the skin until it reaches the peritoneal cavity. [2]By approaching the epidural space from the midline, below or above the scar, we minimized the risk for contact of the Tuohy needle with the lumboperitoneal shunt. Tarshis et al. [8]also considered this risk minimal when, without previous imaging, they inserted an epidural needle below the scar in a parturient patient with an implanted intrathecal pump. Although knot formation at the tip of epidural catheters has been reported in the absence of other foreign bodies, it is a rare complication that occurs in approximately 1 in 30,000 cases. [9]Knotting of the epidural catheter with the lumboperitoneal shunt is a theoretical concern. Obviously, the epidural catheter should be removed cautiously. Abnormal resistance during removal should raise the possibility of lumboperitoneal shunt entanglement and dictate appropriate imaging studies.

Ineffective epidural anesthesia has been described in patients who underwent previous back surgery and instrumentation, presumably because of scarring of the epidural space. [10]Lumboperitoneal shunt insertion is a simple surgical procedure that requires minimal instrumentation compared with these more extensive surgical procedures (Harrington rod, spinal fusion). The first catheter provided excellent labor analgesia for 7 h. After transfer of this morbidly obese parturient patient to the delivery room, the same catheter failed to produce surgical anesthesia. Most likely this was a result of catheter dislodgment. Postoperative scarring of the epidural space from previous shunt placement would have been expected to impair labor analgesia. The catheter was replaced and the new one provided excellent surgical anesthesia for cesarean section. Epidural catheter replacement led to administration of a higher than usual total dose of lidocaine. Epinephrine minimized systemic absorption, but this patient's morbid obesity permitted a higher total dose. Nevertheless, divided administration and close monitoring for early signs of local anesthetic neurotoxicity were the keys to safe management of this patient.

With the potential for cesarean delivery, the risk for airway catastrophe was a major factor in electing early neuraxial anesthetic intervention in this morbidly obese parturient patient with preeclampsia. The benefits of epidural anesthesia were judged to outweigh the theoretical risks to the lumboperitoneal shunt. Despite the need for catheter replacement, epidural anesthesia was safely and successfully administered in this patient with a lumboperitoneal shunt.