Febrile Reaction to Subarachnoid Baclofen Administration

A 33-yr-old woman underwent thoracoscopic-assisted T6–T7 discectomy and fusion with autologous rib bone grafting in February, 1998. Postoperatively, retropulsion of the bone graft resulted in spinal cord compression. Surgical reexploration was performed, and the bone graft was removed, but she continued to show signs of myelopathy with a gait disturbance. Magnetic resonance imaging showed syringomyelia extending from T5 to T7 in the spinal cord, and in August 1999, a T5–T6 laminectomy was performed for drainage of the syrinx. Unfortunately, her symptoms persisted after the third surgery. Subsequent imaging studies showed adequate decompression of the syrinx but also myelomalacia in the left lateral portion of the cord.

The patient presented to the Pain Medicine Center with findings reminiscent of a Brown-Sequard syndrome. There was loss of pain and temperature sensation below the T6 dermatome on the right with spasticity and subjective weakness in the left lower extremity. She also reported a burning dysesthesia below T6 on the left, refractory to multiple medications. Attempts to treat her spasticity with oral medications (including diazepam, baclofen, and tizanidine) resulted in excessive negative effects.

The patient was admitted for an inpatient trial of subarachnoid baclofen administration during which she received a single bolus dose of 50 μg baclofen via  lumbar puncture. This dosage resulted in complete resolution of her spasticity for nearly 24 h. However, a fever developed immediately with a maximum temperature of 39.0°C (measured orally) within 2 h of injection. No source of fever was determined. The patient experienced flu-like symptoms, and a child was concurrently at home with the flu. An allergist concluded that the fever might represent an idiosyncratic reaction to the subarachnoid injection of baclofen, although flu was more likely. If the origin of the fever was indeed related to baclofen, it was hypothesized that the fever was unlikely to recur with slow, gradual, continuous dosing (as delivered via  a pump) in lieu of a single “large” bolus. Moreover, her response to subarachnoid administration of baclofen was substantive, and the potential for improved quality of life was significant. She also underwent a successful, separate, outpatient trial of spinal cord stimulation as treatment for her dysesthesia.

Combined implantation of an InDura^®intraspinal catheter with SynchroMed^®EL pump and a Pisces Quad^®epidural neurostimulator lead with Itrel^®3 implantable pulse generator (Medtronic) was subsequently performed at a single operation. The SynchroMed^®pump was filled with 500 μg/ml baclofen before implantation. It was programmed to deliver a single, calculated bolus of 232 μg (to fill the total catheter and pump tubing volume of 0.46 ml) and to run thereafter in a simple continuous infusion mode at a rate of 2.08 μg/h (50 μg/24 h). Twelve hours later, the patient's temperature suddenly spiked to 40.4°C. She also reported bilateral lower extremity weakness. The subarachnoid baclofen infusion was immediately stopped, and a full workup for the source of the fever was initiated. Within several hours, she had defervesced, and her lower extremity strength had improved significantly.

By the next morning, her neurologic examination results had returned to baseline, and another attempt was made to restart the subarachnoid baclofen infusion. The baclofen was removed from the pump reservoir and replaced with a lower concentration of 250 μg/ml. The side port of the pump was aspirated to remove the concentrated baclofen solution (500 μg/ml) from the catheter and to replace it with cerebrospinal fluid (CSF). The pump was programmed to deliver a bolus of 116 μg over 21 min to refill the catheter volume with baclofen. (Because the internal pump tubing has a volume of 0.26 ml that still contained 500 μg/ml baclofen, the patient actually received a subarachnoid bolus of 14 μg. Customarily, a “bridging bolus” is performed to correct for the internal pump tubing, so only the length of the catheter is refilled [without administration of a bolus] before initiation of the infusion. In this situation, it would have required 221 h to complete.) A simple continuous infusion was programmed to run at a rate of 1.04 μg/h (25 μg/24 h). Two hours after the infusion was begun, lower extremity weakness and fever with a maximum temperature of 38.8°C again developed in the patient. The infusion was stopped, and the patient rapidly defervesced and regained lower extremity spasticity.

During her hospital stay, the patient's leukocyte count remained within the normal range, varying from 6,000 to 10,100/μl, with a normal differential. Erythrocyte sedimentation rate obtained on the third postoperative day was increased at 57 mm/h. No infectious source for the fever could be identified. Mammography and abdominopelvic computed tomography revealed no evidence of malignancy. CA-125 tumor marker was negative. Chest radiography and subsequent computed tomography showed bilateral hilar adenopathy. Skin testing revealed anergy. Angiotensin-converting enzyme concentrations were increased, and sarcoidosis was thought to be the likely etiology of the radiographic findings. Definitive histologic diagnosis was made by bronchoscopy after discharge from the hospital.

The subarachnoid baclofen infusion was halted from postoperative day 3 through postoperative day 6 during diagnostic evaluation of the lymphadenopathy. During this period, the patient remained afebrile, with a maximum temperature of 37.8°C. On postoperative day 7, another attempt was made to restart the subarachnoid baclofen infusion at a rate of 1.04 μg/h (25 μg/24 h). A fever to 38.5°C developed 7 h after the start of the infusion, and the infusion was again halted. The patient was afebrile overnight and was discharged to home the next morning.

Fourteen months after implantation, the patient continued to have significant relief of her dysesthesia with spinal cord stimulation. Subarachnoid administration of baclofen was never resumed.

Fever or hyperthermia has been described to occur as a result of withdrawal from intrathecal baclofen. 8,9Several cases have been reported of a syndrome resembling malignant hyperthermia or neuroleptic malignant syndrome. The manifestations of baclofen withdrawal syndrome include severe spasticity, high fever, hypotension, mental status changes, and rhabdomyolysis. 8,9 

There have been no previous published case reports of hyperthermia occurring in association with the administration of subarachnoid baclofen by bolus or continuous infusion. On four separate occasions, fever with maximum temperatures ranging from 38.5 to 40.4°C consistently developed in this patient after administration of subarachnoid baclofen by bolus injection or continuous infusion, thereby suggesting causation.

Baclofen has been reported to cause marked increases of body temperature in the rat when injected directly into the cerebral ventricles. 10,11If such thermoregulatory physiology was operant in this patient, a potential mechanism underlying the genesis of fever could be the spread of baclofen within the CSF to higher rostral centers. Longitudinal spread along the neuraxis after drug administration is dependent on (1) diffusion into nervous tissue, (2) uptake into blood vessels (both are facilitated by enhanced lipid solubility), and (3) CSF bulk flow. Baclofen is hydrophilic, allowing greater distribution along the neuraxis via  rostral CSF bulk flow.

There was a delay in the onset of fever associated with all baclofen infusions. The onset of fever in each case corresponds with the time required for a hydrophilic agent to reach the ventricular systems via  CSF bulk flow (3–12 h). 12,13The second febrile episode was associated with a 14-μg subarachnoid bolus of baclofen, which may explain the hastened onset of fever (2 h). The rapid injection (seconds) of a “large” bolus dose (50 μg) at time of trial may have facilitated both longitudinal spread within the neuraxis and systemic uptake, explaining the rapid onset of fever. There is no known association between baclofen and sarcoidosis with respect to the genesis of fever.

The decision to proceed with implantation was obfuscated by the patient's flu-like symptoms at the time of trial and the reassurance of the allergy consultation. In the case of fever with undetermined etiology at the time of trial screening in an otherwise healthy individual, it may be wise to repeat the trial with percutaneous subarachnoid catheterization in lieu of bolus injection. Incremental stepped infusions of baclofen may be administered at 12-h intervals (i.e. , the time for a hydrophilic drug to reach higher rostral centers via  CSF bulk flow). With the use of stepped, increasing infusion rates for baclofen administration, (1) the presence or absence of a causal link between baclofen and fever can possibly be shown; (2) a threshold dose of baclofen that must be reached to achieve fever can possibly be shown (although not demonstrated in this patient, such a pharmacologic relation is possible); or (3) the absence of fever can reaffirm that slow, gradual, continuous dosing can avert fever in lieu of a single “large” bolus, if fever is generated by a subsequent bolus injection through the catheter. (If the hypothesized mechanism of fever genesis via  CSF bulk flow due to the hydrophilic character of baclofen is correct, the third possibility will not occur.)

Continuous subarachnoid administration of baclofen to a spastic patient can represent a dramatic improvement in his or her quality of life and is therefore difficult to withhold in the case of a successful trial. In rare situations in which baclofen administration may be causally linked to fever, percutaneous subarachnoid catheterization with incremental, increasing, stepped infusions of baclofen may facilitate the decision to proceed or to abort implantation.