Baclofen bridging, weaning protocol and pain management of a person with T6 paraplegia who required removal of intrathecal baclofen pump due to wound infection

Abstract

Intrathecal baclofen is recognised as an effective treatment option for severe symptoms of segmental spasticity after spinal cord injury. We present our experience of a case of a patient who was admitted to our centre for elective intrathecal baclofen pump revision surgery, but postsurgical complications of a blocked catheter and subsequent wound infections meant the pump had to be removed. In this case, treatment of severe spasticity and nociceptive pain was required while avoiding potentially lethal side effects of sudden baclofen withdrawal. We suggest an effective bridging and weaning protocol for oral drug titration to treat spasticity and pain as a blended syndrome in this challenging situation.

Background

Intrathecal baclofen (ITB) is recognised to be very effective in severe, segmental spasticity after spinal cord injury (SCI) and has good protective effects against occurrence of autonomic storming.¹ The pump, usually spherical, is surgically inserted under the skin in the abdomen and is connected to a catheter system, which delivers the baclofen into the intrathecal space. The dose must be topped up and adjusted in a specialist secondary-care clinic roughly every 3 months via the abdominal device. The lifespan of an intrathecal pump is usually 5–7 years, when the usual reason for revision is battery failure. There is very little published guidance or research on bridging or weaning protocols of baclofen in the event of pump failure/complications. Sudden baclofen withdrawal can cause a potentially lethal syndrome, including rebound of severe spasticity, painful spasms, seizures, agitation, insomnia, confusion, hallucinations, dyskinesia and psychosis; therefore, it is imperative to prevent this through a planned weaning pathway of the ITB and bridging with oral antispasmodics and analgesia.²

This report focuses on our experience of a case from our spinal rehabilitation centre where multimodal treatment of severe spasticity and nociceptive pain was required when an ITB pump was removed. We suggest a baclofen bridging and weaning protocol and special considerations for these patients in the immediate postoperative setting. At the severely symptomatic end of the spectrum, pain and spasticity are often experienced as one and the same syndrome, often inextricably linked in the patient’s subjective experience.

Case presentation

A 48-year-old man sustained a traumatic SCI in 2010 after falling approximately 40 ft from a window. He was previously well with no significant previous medical history. He suffered a T8 vertebral fracture and C5/6 intervertebral disc displacement with cord compression. He underwent surgical fixation at a major London hospital. As there was no sensation below the level of the injury, his injury was classified as T6 complete paraplegia.

During the forthcoming 3 years as an outpatient, he reported ongoing severe spasticity characterised by painful leg spasms frequent over a 24 hour period. His SCI was reclassified as incomplete as he had recovered sensation to deep anal pressure. The spasm frequency and magnitude were unresponsive to conservative measures (physiotherapy and mindfulness interventions) and administration of oral antispasticity management including baclofen, tizanidine and benzodiazepines. Interestingly, he described his neuropathic pain perception at this stage as well controlled, although he was not treated with gabapentinoids or opioids. A trial of amitriptyline did not reduce his perception of discomfort from spasms.

Treatment

ITB treatments and revisions

The patient had his first pump inserted in 2013; a Medtronic SynchroMed II pump. This greatly improved his spasms for the following 6 years and he enjoyed a period of stable symptoms.

In 2020, he underwent routine elective revision of the baclofen pump due to predictable battery failure. No intraoperative complications were experienced and he was discharged for outpatient follow-up. After several weeks, he contacted the rehabilitation centre with worsening painful abdominal and lower limb spasms, but no other medical concerns or noxious stimuli identified. He was readmitted under the rehabilitation team 6 weeks after his surgery, for a multidisciplinary review into possible causes.

A trial of a higher bolus dose of ITB was administered—an increase in 20% original dose (374.6 μg/24 hours from starting dose 309.3 μg/24 hours). However, there was no change in either spasms or pain perception. It was concluded that the catheter was blocked and he underwent a surgical pump replacement (16 June 2020). Flucloxacillin and gentamicin were used as prophylactic antibiotics on this occasion.

After discharge, he developed a surgical wound infection requiring readmission for the third time for a wound washout and debridement. Prophylactic gentamicin was used prior to surgery. Tissue cultures grew Staph epidermis and Corynebacterium amycolatum. Intravenous teicoplanin was started before switching to oral linezolid for a further 2 weeks (3 weeks total). He was discharged home.

Three weeks later, he was readmitted a final time after developing a fever and purulent discharge from the anterior abdominal wound. The multidisciplinary team (rehabilitation, surgery and microbiology) agreed that the best solution was to remove the pump. Due to the ongoing infection, it was deemed too risky to insert a new pump. The operation was challenged by the fact that there was significant difficulty removing the internal catheter under tension in the intrathecal space. The external catheter tip was, therefore, ligated and the rest of the catheter artefact remained in situ. Superficial and deep wound tissue cultures grew Staph epidermis, Klebsiella pneumonia, Serratia marcescens and Citrobacter Koseri. He commenced intravenous antibiotics consisting of meropenum (later switched to ertapenum) and teicoplanin. These were converted to oral antibiotics (doxycycline and cotrimoxazole) on his discharge, with the total course being 12 weeks. He currently remains without a pump but is keen to reinstate the device for spasticity relief despite previous complications.

Baclofen bridging and weaning protocol and pain management

Three days prior to surgical pump removal, the ITB dose was decreased by 10% (374.6 μg/24 hours to 337.2 μg/24 hours) and the bridging process was commenced, with oral baclofen at 30 mg three times per day and 250 μg clonazepam nightly for breakthrough spasms (see table 1). The following day, the ITB dose was reduced by a further 20% (to 268.9 μg/24 hours). The dose was set as a rate for a patient who was not previously baclofen-naïve, and he was counselled on the possible side effects of oral treatment such as drowsiness and confusion, dry mouth, nausea and headache. Use of a benzodiazepine had a double action for spasticity and also to bridge the possible reduction in the seizure threshold that would be experienced with the oral dose of baclofen.

Table 1.

Perioperative ITB pump bridging and weaning protocol and pain management

Time to surgery	ITB dose	Oral spasticity agent	Breakthrough spasms/anti-seizure agent	Nociceptive pain agent
3 days prior to pump removal	Decreased by 10%	Baclofen 30 mg three times a day	Clonazepam 250 mcg once nightly
2 days prior to pump removal	Decreased by 20%	Baclofen 30 mg three times a day	Clonazepam 250 mcg once nightly
Day of pump removal		Baclofen 30 mg four times a day +tizanidine 2 mg twice daily	Clonazepam 500 mcg once nightly	Paracetamol 1 g four times daily +tramadol 50 mg four times daily
3 days postsurgery		Baclofen 30 mg four times a day +tizanidine 4 mg four times daily	Clonazepam 250 mcg as required	Paracetamol 1 g four times daily +tramadol 50–100 mg as required

ITB, intrathecal baclofen.

On the day of final pump removal, the frequency of oral baclofen was increased, and clonazepam was increased. Tizanidine was also started at 2 mg two times per day—there would, therefore, be scope to increase oral baclofen and/or tizanidine at a later date. After surgery, a 1% intravenous propofol infusion was started to bridge the patient for potential seizures, starting at 20 mg/hour alongside the oral medications. Tizanidine was increased postoperatively to 4 mg four times per day. The intravenous propofol infusion was gradually titrated over the next 48 hours in increments of 10 mg/hour until a final dose of 60 mg/hour.

While in intensive care, within 12 hours of his operation, his spasms became more severe. Although he was not reporting explicitly of pain at this stage, he was prescribed regular analgesia postoperatively on his drug chart, according to the NICE guidelines on perioperative care in adults.³ Paracetamol and tramadol were administered—the dose was rapidly increased from 50 mg as required to 100 mg regularly four times per day. The nociceptive pain and spasms became much more manageable and sleep was no longer elusive. Strong opiates were never required. The patient’s spasms became more controlled and the clonazepam was reduced and used when required. Two days later, the patient was rapidly weaned off the propofol infusion.

Outcome and follow-up

Progress of rehabilitation postpump removal

On return to the rehabilitation unit, positional factors remained important triggers for painful spasms. Lying flat with hips extended tended to trigger abdominal and hip flexor spasms, which were very uncomfortable. These were significantly reduced with the use of a T-roll support and a passive stretching programme to the lower limbs with feet in full dorsiflexion. As his surgical wounds healed and his immediate nociceptive pain subsided with tramadol, he self-directed a weaning programme for this. The patient found that tramadol caused a subjective slowing of mental faculties, a side-effect he found unacceptable. As there were no neuropathic features of this pain, he did not require additional gabapentinoids.

Focal spasticity to the hip flexors was treated with botulinum toxin A (Dysport is used in this centre) to his hip flexors bilaterally. Any interface with neurogenic bladder spasms or visceral discomfort was treated with oxybutynin.

On discharge, his only medication from the rehabilitation unit was tramadol 50 mg as required, which he continued to wean. Clonazepam was only occasionally used as required for breakthrough spasms—the frequency of which continued to reduce. The multidisciplinary inpatient treatment plan was so successful that the team plan to wean tizanidine orally in due course as an outpatient, according to the patient’s reported severity of spasticity while at home. Liaison with the spinal surgical team was ongoing regarding installation of a new baclofen pump, a procedure the patient is keen to undergo.

Discussion

Taking a baclofen pump out suddenly can be detrimental to the patient, and there is a significant lack of agreed guidelines for undertaking this for clinicians. Our case highlights that although this is a rare event, it is important to be aware of the serious side effects of sudden cessation of ITB.⁴ These include confusion, hallucinations, new onset seizures and in extreme circumstances, death. For our patient, avoiding these side effects provided the rationale for the propofol infusion administration in the immediate postoperative period after pump removal. An alternative treatment option to negate the regular monitoring required for baclofen dosing would be intrathecal phenol. However, this approach would result in the permanent loss of bladder, bowel and limb control below the level of the lesion.⁵ This option was not attractive to our patient as his SCI classification had improved since the original injury—he started as a complete paraplegic SCI after his injury (no sensation below the level of the injury) and progressed to an incomplete injury, meaning some sensation was recovered below the level of injury.

One novel approach by Hwang et al used an externalised ITB pump in order to slowly wean patients off the baclofen pump, which was infected.⁶ They successfully managed to avoid any withdrawal symptoms in patients, other than the expected mild increase in spasticity. Using an externalised system consisting of the explanted pump, it allowed them to reduce the infusion rate of baclofen by 20% while simultaneously increasing oral baclofen and tizanidine. Each subsequent day, the ITB infusion was decreased by 20%–50%. We did not have capacity to reproduce this novel procedure at our centre.

A circumspect approach was the best tool to get the potential triggers of spasms under control. Triggers included the presence of nociceptive wound pain from surgery, musculoskeletal back pain from the ITB insertion site and a neurogenic bladder. Getting the spasms under control using standard, ubiquitous pain agents in combination with traditional ‘anti-spasticity’ agents have mitigated the use of additional opioid and neuropathic pain medications later on.

A serious adverse event of a surgical wound infection unfortunately occurred, despite preventative infection control measures taking place according to the NICE guidelines: surgical site infections: prevention and treatment 2019.⁷ The wound infection was treated successfully according to the local antimicrobial guidelines and advice was sought immediately from the microbiology team. Multidisciplinary team input and discussion meant that the patient made a full recovery after the pump removal and after completing the course of antibiotics. They did not need to return to hospital other than the routine check-ups.

The case illustrates our approach to baclofen bridging and weaning after ITB pump failure. It also highlights that the experience of severe spasticity can be considered by patients to be a pain syndrome in its own right rather than two mutually exclusive entities, requiring a more blended approach to pharmacological management. However, we acknowledge that the qualitative/anecdotal reporting of pain perception is a limitation of this report; the use of objective pain scores throughout the experience, such as the Numeric Graphic Rating Scale or Scale of Pain Intensity (SPIN) screen would have been beneficial and a more ‘robust’ temporal record of pain perception.⁸

Learning points.

• Patients who have their intrathecal baclofen pump removed must be closely monitored with an oral replacement bridging protocol in order to avoid the side effects of sudden baclofen withdrawal. Monitoring should include presence of spasticity and pain symptoms and treatment should address both of these issues.

• Patients who have undergone spinal surgery need to be meticulously reviewed for noxious stimuli, irrespective of the severity of this patient’s original spinal cord injury.

• Pain in this subset of patients must be evaluated as a differential of nociceptive and/or somatic and/or visceral origins and treated promptly as such.

Ethics statements

Patient consent for publication

Obtained.