MUSCLE spasticity is a feature of cerebral palsy that is due to simultaneous contraction of agonist and antagonist muscle groups. Descending inhibitory interneurons are affected, causing an excess of excitatory neurotransmitters and inadequate release of γ-aminobutyric acid (GABA) in the spinal cord. Overstimulation of α motor neurons results in these contractions. Baclofen (Novartis Pharma AG, Basel, Switzerland) acts as an agonist at the GABAbreceptors in the dorsal horn of the spinal cord and reduces the tone and pain associated with muscle spasms. 1Intrathecal administration of the drug reduces spasticity at lower doses than are needed orally and with fewer side effects. These side effects include bradycardia and hypotension during general anesthesia and delayed arousal after surgery, all of which have been reported in patients who were taking oral baclofen preoperatively. 2,3We describe a case in which these features occurred in association with the intrathecal administration of baclofen.
Case Report
A 9-yr-old, 19.4-kg child with cerebral palsy was admitted with a history of fluid accumulation over a SynchroMed® (Medtronic, Minneapolis, MN) baclofen infusion pump. The pump had been subfascially implanted 3 yr prior to this event. The catheter entered the intrathecal space in the lumbar region, and its tip was located in the mid to upper thoracic region. An abdominal roentgenogram showed a fracture of the catheter at its hub connection to the port of the pump. The patient was alert, with a history of increased muscle spasms. His other medications included oral ranitidine and transdermal scopolamine for control of gastroesophageal reflux and nausea. The patient's blood pressure and heart rate were 113/85 mmHg and 89 beats/min, respectively. No premedication was given.
Anesthesia was induced with 100 mg thiopental and 50 μg fentanyl. Rocuronium, 20 mg, was used to facilitate tracheal intubation. Anesthesia was initially maintained with 0.5% isoflurane and 70% nitrous oxide in oxygen. The pump was programmed to stop infusing baclofen (2,000 μg/ml) shortly after induction. Over the next 30 min, the patient's blood pressure and heart rate were noted to decrease to 55–60/30–40 mmHg and 55 beats/min, respectively, despite discontinuation of the inhalational agent and the onset of surgery. Volume administration and intravenous glycopyrrolate, 0.1 mg, increased the heart rate to 100 beats/min and increased the blood pressure to 80/45 mmHg. The oxygen saturation remained at 98–100% throughout the procedure. The fractured hub connector was removed and replaced with a two-piece catheter that was trimmed and spliced to the existing catheter. Residual baclofen solution and cerebrospinal fluid was allowed to drain back through the repaired catheter before it was attached to the pump. The pump was then reprogrammed to deliver a bolus of 777 μg baclofen (0.3 ml) to prime the catheter and then to infuse at a rate of 15.6 μg/h. At the end of the procedure, which lasted 1 h, we noticed that the patient was unresponsive despite adequate reversal of neuromuscular blockade. The pupils were 2 mm in size. Intravenous naloxone was given in 20-μg increments to a total of 120 μg. Spontaneous ventilation and airway reflexes subsequently returned. When the patient started to awaken, he was extubated and transferred to the recovery room, with a blood pressure of 100/70 mmHg, a heart rate of 100 beats/min, and a respiratory rate of 20 breaths/min.
The patient then became more somnolent in the recovery room, despite gentle stimulation. His blood pressure and heart rate gradually decreased to 75/31 mmHg and 59–62 beats/min, respectively, 70 min after the pump was activated. His respiratory rate was 10 breaths/min, and his oxygen saturation, as measured by pulse oximetry, was 100% on 30% inhaled oxygen. Venous blood glucose concentration was 90 mg/dl. No improvement in consciousness level occurred after administration of 5 g dextrose or after further administration of naloxone. The pupils were midsized. Volume administration and atropine, 0.3 mg, increased the heart rate to 80–90 beats/min and increased the blood pressure to 80/40 mmHg. At this time, it was realized that the administered bolus dose of baclofen may have been greater than was necessary to prime the catheter. The pump was turned off, and the patient, still unconscious but breathing at 15–20 breaths/min, was admitted overnight to the pediatric intensive care unit for observation. Two further doses of atropine, 0.3 mg, were given to maintain the heart rate in the 80–90 beats/min range.
Twelve hours after the bolus dose was given, the patient began to awaken. The intrathecal baclofen infusion was resumed at a rate of 375 μg/24 h. However, the patient again became unconscious 4 h after starting at this rate, so the infusion was stopped again. When consciousness returned, the infusion was started at 99.8 μg/24 h, with no further recurrence of coma. The patient's blood pressure and heart rate returned to baseline values. A subsequent increase in the infusion rate to 150 μg/24 h also did not affect consciousness. The patient was discharged in stable condition at a final rate of 250 μg/24 h on the third postoperative day.
Discussion
There are two major indications for the use of intrathecal baclofen in patients with cerebral palsy. The first is to improve the ability of patients with spastic diplegia or quadriplegia to ambulate with or without external aids. The second indication is a reduction of tone in nonambulatory patients with severe spasticity to expedite their nursing care. 4At a preoperative screening trial, bolus doses of 25–100 μg of the drug are given intrathecally to assess their effect on muscle tone (Ashworth scale). The pump is later implanted if an improvement in tone is demonstrated. The pump is programmed with an external computer to deliver a bolus of drug, which fills the dead space of the pump and the intrathecal catheter, and then to provide a continuous infusion at a predetermined daily dosage. The priming bolus is calculated from the drug concentration, the internal pump tubing volume, and the implanted catheter volume, which is dependent on its length. The daily dosage is adjusted periodically to achieve the most optimal effect on tone. This dosage is not related to age or weight. 5A large variation in dose, ranging from 22 to 869 μg/24 h, has been reported. 4
Intrathecal infusion of baclofen results in much higher concentrations of the drug in cerebrospinal fluid than are achieved with oral administration; thus, there is a greater reduction in spastic tone and a lower incidence of cerebral side effects. 6When the intrathecal catheter tip is located in a midthoracic rather than a lower thoracic position, there is a greater improvement in upper extremity tone, without a decrease in the drug's effect on lower extremity tone, in children with quadriparetic spasticity. The midthoracic position also allows a reduction in daily dosages to achieve this effect because of the more even distribution of the drug in the cervical and lumbar cerebrospinal fluid. 7Intrathecal baclofen has a half-life of 5 h and a clearance rate of 30 ml/h when it is given either as a bolus or as an infusion. 4
The anesthetic agents potentiated the central nervous system effects of baclofen to cause hypotension and bradycardia during surgery. In awake patients, baclofen does not have pronounced cardiovascular effects, although a mild hypotensive action occasionally is seen. 8In anesthetized animals, GABA and drugs that stimulate GABA receptors cause a decrease in arterial pressure and heart rate. 9Baclofen, which acts in part by stimulating central nervous system GABA receptors, may be synergistic with general anesthetics that potentiate GABA action on synaptic transmission. The result is a reduction in sympathetic outflow to the vasculature and to the heart. 10The effect may be augmented by fentanyl, which stimulates central vagal nuclei. 11In addition, intrathecal baclofen is antinociceptive in several animal models. 12GABAbagonists may regulate the excitability of the primary afferent terminal in the dorsal gray matter of the spinal cord by reducing the release of transmitters from these terminals and by causing postsynaptic hyperpolarization. Spinal baclofen produces a dose-dependent inhibition of the C fiber–evoked activity in dorsal wide–dynamic range neurons. 13In contrast to animal studies, a comparable analgesic effect of intrathecal baclofen has not, however, been consistently reported in humans. 14
The effect of the baclofen bolus dose was superimposed on the increased sensitivity to anesthetic agents to cause unconsciousness to persist in the postoperative period. Bolus administration is more likely than infusion of the intrathecal drug to be associated with adverse events, such as sedation, bradycardia, hypotension, and respiratory depression. The side effects depend on the degree of the overdose, which determines the amount of drug that reaches and affects supraspinal receptors. 15Bolus administration at the upper to midthoracic level may cause earlier supraspinal diffusion and effects of the drug. In contrast to slow infusion, the initial drug distribution is influenced by the drug volume and density during bolus injection. 16The density of the 2,000-μg/ml baclofen solution was determined to be 1.003 g/ml at 24°C. The value is similar to that previously obtained for the 500-μg/ml solution at the same temperature. 17The specific gravity of the latter solution at 37°C was reported to be 0.9996. The specific gravity of the 2,000-μg/ml solution is likely to be similar at body temperature. Thus, the solution, which is made up in normal saline, is isobaric or slightly hypobaric in relation to cerebrospinal fluid. The subsequent distribution of the drug is determined by its hydrophilic nature. Clearance rates obtained from bolus and infusion studies are similar, indicating that baclofen is cleared by rostral bulk flow of cerebrospinal fluid. 4
Overdoses are more likely to be caused by human error in the programming or refill procedure than by pump malfunction. 1The use of a high-concentration solution causes an overdose to occur easily because of the small priming volume. In our case, the calculated priming bolus was appropriate to fill the dead space of the pump and the catheter. As the pump was undisturbed, its dead space was already filled with baclofen, so that a smaller dose (0.1–0.15 ml) was needed to fill the catheter. Thus, about 0.2 ml (400 μg) was administered by bolus into the spinal fluid. Alternative factors that cause delayed awakening after general anesthesia, such as other medications, metabolic disturbances, and neurologic injury, were excluded. 3Naloxone was given to fully reverse the sedative and respiratory depressant effects of fentanyl. Blood glucose and electrolyte concentrations were normal. Systemic uptake and redistribution to the brain were considered unlikely causes of early postoperative coma, although plasma baclofen concentrations were not measured. Systemic absorption of the high-concentration drug would most likely have occurred from the fluid pocket around the pump during surgery, as baclofen administered intrathecally is not associated with significant concentrations in the systemic circulation. 4The preoperative awake status of the patient suggests that plasma concentrations resulting from any absorption were insufficient to affect the consciousness level. The lack of increased spasticity in the upper extremities in the early postoperative period indicated that there was rostral migration of the bolus dose to the cervical and supraspinal intrathecal areas, which was facilitated by the high spinal position of the catheter. The recurrence of coma, after initiation of the intrathecal infusion at the previous rate on the first postoperative day, may have indicated increased receptor sensitivity secondary to the decreased intrathecal baclofen concentrations caused preoperatively by leakage from the fractured catheter. 18Postoperative reduction in the infusion rate was therefore required.
Preoperative discontinuation of oral baclofen therapy has been advised to prevent the occurrence of severe hemodynamic instability during anesthesia. 2However, similar discontinuation of intrathecal therapy may cause an increase in tone and muscle spasms before surgery in patients with cerebral palsy. Rapid withdrawal of intrathecal baclofen after chronic administration also has been associated with seizures, psychosis, hallucinations, and visual disturbances. 19Ephedrine, atropine, and epinephrine all have been reported to be effective in treating baclofen-induced hypotension and bradycardia during general anesthesia. 2The emergency protocol, recommended in the case of an intrathecal baclofen overdose that causes coma and respiratory depression, advises maintaining the airway, breathing, and circulation; terminating the pump infusion (if necessary, by emptying the pump); and administering physostigmine, 0.02 mg/kg, at a rate of no more than 0.5 mg/min. The dose is repeated at 5–10-min intervals until a therapeutic effect is obtained or a maximum dose of 2 mg is given. 20The drug was not given in our case. We decided to allow the patient to recover spontaneously with airway and circulatory support, as long as spontaneous ventilation and oxygenation were maintained. Lumbar puncture and withdrawal of 30–40 ml of cerebrospinal fluid, which can be assayed to show an abnormally high concentration, has also been used to reverse unconsciousness. 4
In conclusion, intrathecal baclofen overdose should be considered as a possible cause of hemodynamic instability during pump or catheter revision procedures or of delayed emergence in the postoperative period. Increased sensitivity to anesthetic and analgesic agents warrants caution in the use of these drugs during surgery. Appropriate therapy should be readily available to treat cardiovascular, neurologic, and respiratory disturbances that may occur in association with the intrathecal medication.