Patient Selection

Candidates for ITB usually have spastic quadriparesis with mean Ashworth scores of 3 or greater in the upper and lower extremities (Table 226-1). Most have had no significant improvement with oral medications and botulinum toxin injections, but ITB is occasionally appropriate for children with severe spastic quadriparesis who have had no previous treatments because the likelihood of sufficiently improving severe spasticity with oral medications is low. Their spasticity is either impeding caregiving, causing progressive contractures, interfering with function, or causing pain. ITB is used most often in children who weigh 15 kg or more and are 4 years or older, but with the smaller pumps that are now available, there are virtually no age or size limitations. Palisano and coworkers developed the Gross Motor Function Classification System (GMFCS) to broadly categorize levels of disability of children with CP.⁷ Most children who are candidates for ITB are GMFCS level IV or V, occasionally level III, and rarely level II.

TABLE 226-1 Ashworth and Modified Ashworth Scales

It is important that the patient and caregiver goals with ITB therapy be reasonable and clearly identified preoperatively. “Reducing spasticity” per se is not a therapeutic goal. Appropriate goals include increasing range of motion, facilitating care, slowing the development of musculoskeletal contractures, and decreasing painful muscle spasms.

ITB is rarely administered with the intent of improving gait in patients with CP who have spastic diplegia and should not be considered as an alternative to lumbar rhizotomies; the two modalities are indicated for different groups of patients—young ambulatory patients with spastic diplegia are best treated with lumbar rhizotomies, whereas nonambulatory or minimally ambulatory patients with spastic quadriparesis are better treated with ITB. Administration of ITB can be considered in an attempt to improve gait in young adults with spastic diplegia, but in such cases, infusion screening trials are helpful to evaluate the effects of ITB on gait before a pump is implanted.

Screening

When ITB was first used to treat spasticity of cerebral origin, screening trials were conducted to evaluate its effectiveness before pumps were inserted. Screening trials involved the injection of a bolus baclofen dose, 25 to 100 µg via lumbar puncture, and comparison of spasticity scores in the lower extremities before the injection with scores at 2-hour intervals afterward for 8 hours. Decreases of 1 point in the mean lower extremity Ashworth score were considered to be clinically significant and occurred almost invariably. Since then, hundreds of children with spastic CP have undergone screening bolus injections, which virtually always decrease spasticity, and many centers now forego a screening injection. The advantages of a screening trial include (1) the demonstration that baclofen is effective for that patient and (2) the demonstration that function is at times improved by ITB. Disadvantages of screening trials include (1) the cost and discomfort of the trial (headache, nausea, and vomiting are common), (2) the fact that a bolus dose often causes such hypotonia in the lower extremities that pivot transfers and gait are temporarily worsened, and (3) the common development of headache and vomiting from a CSF leak afterward.

If a screening trial is done, however, measurement of opening pressure may be helpful because the pressure is often abnormally increased, and in such cases the risk for CSF leakage is enhanced when the baclofen pump catheter is inserted.⁸ Alternatively, it is reasonable to perform a quick brain magnetic resonance imaging or computed tomographic scan of the head to see whether ventriculomegaly is present; if it is, insertion of an intracranial pressure monitor for 24 hours will provide a better evaluation of intracranial pressure than a single pressure measurement via spinal tap. If pressures are abnormally high, the likelihood of a CSF leak with the baclofen catheter can be diminished either by inserting a ventriculoperitoneal shunt before pump implantation or by inserting the intrathecal catheter with open techniques (laminotomy and purse-string suture around the entry site).

If improving gait is a therapeutic goal, the effects of ITB on gait can be evaluated before pump placement with an intrathecal catheter connected to an external microinfusion pump. Bleyenheuft and coworkers evaluated gait with continuous infusions of baclofen, 45 to 150 µg/day, in nine patients whose gait was deteriorating.⁹ Gait improved in seven of nine as evaluated on the Observation Gait Scale, and the improvement was present after pump implantation.

Pump and Catheter Implantation

Operations to implant a baclofen pump in patients with CP are done under general anesthesia and usually last 1 to 1.5 hours.¹⁰ Programmable pumps are needed when treating patients with spasticity of cerebral origin because of the frequent need to adjust the ITB dosage. At present, the only commercially available programmable pump is the Medtronic SynchroMed pump, a battery-powered pump with a 7-year battery life. The pump is usually implanted in the abdominal wall, subfascially in most children and thin patients and on the fascia in patients with more subcutaneous fat. In patients with no abdominal site for pump implantation (e.g., because of multiple scars, ostomy sites), the infraclavicular region is an alternative site. The baclofen catheter is connected to the pump and tunneled posteriorly to the lumbar region, where it is inserted via a Tuohy needle into the thecal sac and advanced cephalad to a site that depends on the distribution of the patient’s spasticity: T10-11 for patients with spastic diplegia and C7-T4 for those with spastic quadriparesis. For patients with spine fusions or anticipated spine fusions, catheters can be tunneled obliquely cephalad and inserted into the thecal sac via a small laminectomy at C6-7 above the spine fusion, which usually extends from T1-2 to the sacrum (Fig. 226-1).

FIGURE 226-1 Baclofen pump and the catheter tunneled to enter the thecal sac above the spine fusion.

The initial reports of ITB for the treatment of spinal spasticity associated with spinal cord injury or multiple sclerosis positioned the catheter tip in the T10-12 region. However, Kroin and associates subsequently demonstrated that the concentration of radioisotope administered in the lumbar region was 4 times higher than the concentration at the craniovertebral junction.¹¹ To increase the baclofen concentrations in the cervical region and obtain better reductions in upper extremity spasticity, catheter tips are now positioned higher when treating spastic quadriparesis. Grabb and coauthors reported that higher catheter positions were associated with a better reduction in upper extremity spasticity than was observed with lower catheters.¹² Cervical catheter tip positions do not result in increased complications or adverse side effects and are associated with significantly decreased spasticity in both the upper and lower extremities.

Dosing of Intrathecal Baclofen

After implantation of the pump, ITB doses are titrated over the first few months, regardless of the movement disorder being treated. When treating spasticity, infusions often begin at 100 µg/day and are increased by 10% to 20% daily until spasticity is perceptibly reduced to the patient and physician; doses are then adjusted at less frequent intervals on an outpatient basis. Stable doses are achieved by 6 to 12 months. The average ITB dose at 1 year when treating cerebral spasticity is approximately 300 µg/day. Doses greater than 600 µg/day are uncommonly needed to treat spasticity. Some patients seem to respond better to intermittent baclofen boluses, such as 75 µg every 4 hours, than to continuous infusion, but no studies have been reported in which the effects of continuous infusions were compared with those of intermittent bolus infusions in the same patient.

Tolerance to ITB probably does not develop in people with spastic CP. If patients have been at a stable, effective dose for 2 years but then become more spastic and higher and higher doses do not yield the effectiveness of the initial infusion, there is almost always a malfunction of the infusion system and the problem is generally related to the catheter. Approximately 8% of people with dystonic CP do not achieve satisfactory improvement even at high ITB doses, such as 1500 µg/day. The lack of effect is more likely due to unresponsiveness to ITB than to tolerance.

Outcomes of Intrathecal Baclofen in Patients with Spastic Cerebral Palsy

The efficacy of ITB in reducing spasticity in children and young adults with CP has been documented in multiple publications from both single institutions and multicenter trials.^13,14 Butler and Campbell reviewed the 12 publications on ITB that had been published by 2000 and concluded that spasticity is improved and function is probably improved.¹⁵ Spasticity is significantly decreased in the lower and upper extremities and remains decreased during the years of infusion, probably indefinitely; a multicenter study in 2003 reported sustained decreases in Ashworth scores in the lower and upper extremities during 6 years of follow-up.¹⁴ Although the reduction in spasticity in the upper extremities is typically less than that in the lower extremities (if the catheters tips are at the midthoracic level or lower), Motta and colleagues evaluated the effect of ITB on the upper extremities of 20 patients with quadriparetic CP and a mean age of 11.4 years with the Melbourne Assessment of Unilateral Upper Limb Function scale.¹⁶ Upper extremity spasticity was decreased, P < .05, and range of movement and target accuracy improved significantly.

ITB is associated with many other effects besides decreased spasticity. The effect of ITB on function has been evaluated with the Gross Motor Function Measure (GMFM). Krach and coworkers determined GMFM scores in 31 subjects 4 to 29 years old, most of whom had CP, before ITB and 1 year afterward and found that GMFM scores increased 3.7 to 4.1 points in all age groups; the improvements occurred mainly in the domains of transfers, dressing, and rolling.¹⁷ GMFM scores improved in patients who were at GMFCS level II (6.2) and level V (2.9). This group also questioned 100 patients and caregivers after ITB therapy for 1 year. They reported that positioning improved in 60%, transfers in 58%, dressing in 69%, sleep in 43%, and comfort in 53%; 88% indicated that they would undergo the surgery again despite the fact that 32 complications occurred in 22 subjects.

The effect of ITB on caregiving was evaluated with a caregiver questionnaire by Gooch and colleagues in 80 patients who had received ITB therapy for 1 year.¹⁸ The mean age of the patients was 11 years, and all were GMFCS level IV or V (severely motor impaired). Caregivers selected the treatment goals before therapy; the goals of decreased pain and prevention of worsening deformity were achieved in 91% of the patients, and easier caregiving was achieved in 88%.

The effect of ITB on oral motor function, communication, and nutritional status was evaluated by Bjornson and coworkers in 30 children.¹⁹ Speech improved in 10 (including 5 at GMFCS level V), worsened in 2, and was unchanged in the remainder. Use of assistive technology improved in 6, including 5 at GMFCS level V, for which communication is perhaps the most important function that can be improved. In a single case study, Leary and associates found that speech intelligibility was improved after ITB therapy as evaluated by the Assessment of Intelligibility of Dysarthric Speech Scale.²⁰

The reduction in spasticity with ITB is associated with increased range of motion in the lower extremities and may be accompanied by improved gait. Gerstzen and colleagues graded gait on four functional levels (community, household, nonfunctional, and nonambulatory) in 24 patients with some ambulation before ITB therapy.²¹ The level of ambulation improved by one level in 9, worsened in 3, and was unchanged in the remainder. Hip subluxation, which is related to adductor spasticity and is common in children with CP, appears to be diminished by ITB therapy. Subluxation in 33 children who received ITB for 1 year was either improved (in 5%) or unchanged (in 91%).²²

Intrathecal Baclofen for Dystonic Cerebral Palsy

Dystonic CP can be either focal, hemidystonic, or generalized; the most common form by far is generalized. ITB can be used for either generalized dystonia or hemidystonia but is rarely used for focal dystonia. ITB is probably the treatment of choice for severe, generalized secondary dystonia. If patients do not respond adequately to ITB, deep brain stimulation can be considered. The effects of ITB on dystonia can be graded with either the Barry-Albright Dystonia (BAD) scale, a validated scale that was developed to evaluate secondary dystonia, or the Burke-Fahn-Marsden (BFM) scale, a validated scale that was developed for adults with primary dystonia.^23,24

If screening is desired to evaluate the effects of ITB on dystonia before pump implantation, it can be done by connecting an intrathecal catheter to an external microinfusion pump. Infusion begins at 200 µg/day and increases by 50 µg every 8 to 12 hours until either (1) the mean dystonia score decreases by 25% or greater, (2) adverse side effects such as lethargy develop, or (3) a dose of 900 µg/day is infused without a reduction in dystonia.

Given the postulate that the site of action of ITB when treating dystonia is at the cortical level, catheter tips are often positioned higher, such as at C1-4, than when treating spasticity. Infusion at these levels is not associated with lethargy or respiratory depression. Recently, intraventricular baclofen infusion has been described to treat dystonic CP, particularly in patients with spine fusions or cervical anatomy that make intrathecal implantation difficult.²⁵ Dystonia may respond to intraventricular baclofen at lower doses than needed with intrathecal administration, but insufficient data are available to confirm this observation.

The effects of ITB on children and young adults with severe, generalized dystonic CP were described in a large single-institution study.²⁶ Dystonia scores on the BAD scale decreased significantly within 3 months after ITB therapy began, and the decreases continued during 36 months of follow-up. In that study, ITB infusions through catheters positioned at T6 or higher were associated with BAD scores that were significantly lower (4.9 ± 4.1) than those achieved with catheters positioned at T8 or lower (10.4 ± 7.0). Continued improvement in secondary dystonia with ITB has been observed for up to 11 years. Motta and coauthors recently reported 19 patients with dystonic CP, a mean age of 8.5 years, and GMFCS level V who were treated with ITB and evaluated with both the BAD and BFM scales.²⁷ Dystonia scores on both scales were significantly improved at 3 and 12 months after ITB therapy, and significant improvement occurred in posture and caregiving.

Intrathecal Baclofen for Choreoathetosis in Patients with Cerebral Palsy

There are no published reports of the use of ITB for chorea, athetosis, or choreoathetosis in patients with CP. In my experience, numerous individuals with mixed spasticity and athetosis who were treated with ITB had the expected improvement in their spasticity without appreciable changes in their athetosis. Two individuals with severe, generalized athetoid CP and no discernible spasticity have been treated with ITB; their athetosis improved somewhat, perhaps 30%, but in neither case dramatically. Three individuals with choreiform CP and no discernible spasticity have been treated with ITB with the same result: mild improvement.

Complications of Intrathecal Baclofen for Cerebral Palsy

ITB therapy is associated with frequent complications. Gooch and coworkers reported the complications that occurred in 100 children, primarily with CP, who had undergone 117 pump operations.²⁸ Twenty-four children had 48 complications, most commonly catheter-related problems. Motta and coauthors reported that complications occurred in 31% of 200 patients treated with ITB: 11% had CSF leaks, 7% had catheter problems, 7.5% had infections, and 5.5% had more than one complication.²⁹ Ashworth scores of 3 or greater and age younger than 10 years were associated with an increased risk for complications. Despite the frequent complications that are reported in virtually every series, approximately 90% of individuals with baclofen pumps at the end of battery life prefer to have a new pump implanted and to continue ITB therapy.³⁰

Infections occur in 5% to 10% of patients and are caused most frequently by Staphylococcus aureus. Most infections take place within the first month after pump implantation and are related to bacterial contamination at the time of pump insertion; infections related to pump refills are exceedingly rare. Infections may either be anterior at the pump site—where symptoms include erythema, swelling, tenderness, and fever—or involve the entire system, including the catheter and CSF. Pump infections can sometimes be cured with intravenous antibiotics and surgical cleaning of the pump wound. Infections involving both the pump and CSF generally require removal of the entire system and postoperative antibiotics for 2 to 3 weeks. The optimal time for pump reimplantation after an infection is unknown, but the time of reimplantation has ranged from 6 weeks to 6 months.

CSF leaks occur after baclofen catheter insertion in 5% to 15% of patients with CP (most of whom are children), in contrast to the 3% leak rate reported in adults.³¹ The difference is probably attributable both to the smaller size, thinner tissues, and malnutrition of chronically disabled children and to the presence of higher CSF pressures associated with occult hydrocephalus. Postoperative CSF leaks can be manifested as symptoms of low pressure (postural headache and nausea), accumulation of CSF under the lumbar or abdominal incisions, or both. Mild leaks may respond to bed rest; others require injection of epidural blood at the catheter entry site (blood patches), insertion of lumbar drains, or rarely, open exploration and insertion of a purse-string suture and fibrin glue around the catheter exit site.

As catheter technology has improved, the frequency of catheter malfunctions has decreased from 25% to about 5% to 10%. Malfunctions include disconnection of the catheter from the pump or a leak at the site of catheter connection to the pump, development of a leak through the catheter wall, migration of the catheter out of the thecal sac, and migration of the catheter tip through the arachnoid so that it lies in—and infuses into—the subdural space. Catheter malfunctions can be evaluated by aspirating CSF through the catheter via the pump side port, by injecting contrast material through the catheter and observing its passage with fluoroscopy (Fig. 226-2), by computed tomography around the catheter tip after the instillation of contrast medium, and by injection of radioisotope into the pump to identify small leaks that may not be visible after bolus injection of contrast material.

FIGURE 226-2 Fluoroscopic appearance of an intrathecal catheter injected with contrast material demonstrating extravasation of contrast outside the catheter (arrow).

The relationship between ITB and scoliosis was recently clarified. Although several small series reported previously that ITB caused more rapid progression of scoliosis than was observed before pump implantation, those series had no controls. Severan and associates evaluated 107 children treated with ITB for 2 or more years and found that progression of scoliosis developed or was present in 26 of them.³² These 26 patients were matched by age, gender, and GMFCS level with 26 controls who did not receive ITB therapy. Average curve progression after ITB was 16.3 degrees per year, and in the control group it was 16.1 degrees per year. Scoliosis developed in 21% of 57 patients after pump implantation during 3.6 years of follow-up and in 32% of control patients by the time of maturity.

The effect of ITB on seizures has also been clarified. Authors of small series had reported either increased seizure frequency or no change in frequency. In a substantive study, Bounaguro and colleagues evaluated seizures in 150 children undergoing ITB therapy, 40% of whom had seizures before ITB.³³ After ITB therapy, seizures decreased in 13% and increased in 2% of the children who had seizures before ITB. In children who had no seizures before ITB, only 1 child had a first seizure after ITB.

Most ITB overdoses are iatrogenic and related to programming errors, which can occur when initially filling the baclofen catheter or when changing baclofen concentrations. Mild overdoses cause only listlessness and do not need to be treated. Moderate overdoses cause lethargy and hypotonia and are treated by monitoring pulse oximetry while waiting for the baclofen to metabolize. Some authors have recommended intravenous injection of 1 to 2 mg of physostigmine to treat overdoses, but the effects of such injections are minimal and brief. Severe overdoses cause bradypnea and coma and are treated by assisted ventilation. Spinal taps to barbotage CSF may also help with severe overdoses: 20 mL of CSF is withdrawn and replaced with 20 mL of a 5% dextrose–0.25% normal saline solution 2 or 3 times. Baclofen is not neurotoxic, and after the excess baclofen clears, the infusion—and results—typically resume unchanged.

Abrupt cessation of ITB infusion is followed within a few hours by symptoms of withdrawal, including generalized itching, increased spasticity, and agitation.³⁴ Symptoms range in severity from mild to severe. Mild withdrawal may not be recognized by the patient and the interruption in therapy may go undetected for several months. Severe withdrawal is a serious condition and must be treated urgently. In its extreme form it is associated with fever (106°F), rhabdomyolysis, sometimes multisystem organ failure, and rarely death. Moderate withdrawal is treated with high doses of oral baclofen (e.g., 10 to 20 mg every 4 hours) and may be supplemented with intravenous benzodiazepines (diazepam, 2 to 5 mg every 6 hours) or cyproheptadine (6 mg every 6 hours for 24 hours).^35,36 Bolus baclofen injections via lumbar puncture provide additional improvement until the malfunction can be corrected surgically. Withdrawal is most often the result of malfunction of the catheter, such as disconnection of the catheter from the pump, development of a leak in the catheter, or migration of the catheter out of the thecal sac.

Conclusion

ITB is an excellent method of treating spastic quadriparetic CP, particularly in patients who have some function or potential function. It is also the treatment of choice for severe, generalized secondary dystonic CP. Its effects on choreoathetoid CP are relatively limited. ITB is associated with frequent and sometimes serious complications and is expensive, but it is often the best therapeutic option available for these patients.