INTRATHECAL drug infusions are being used more frequently to treat spasticity and chronic pain. More than 95,000 intrathecal drug administration devices have been implanted in the United States since their development in the 1980s.∥When opioids are part of the intrathecal infusion solution, catheter tip granulomas are emerging as an important potential complication. More than 90 cases of catheter tip granuloma have been reported since 1992.1Complication reporting is currently on a voluntary basis, and this number is likely understated. The granuloma is typically a collection of acute and chronic inflammatory cells derived from the arachnoid layer. This reactive tissue does not seem to directly involve the neural parenchyma1but has the potential to compress the spinal cord and produce permanent neurologic deficits, including paralysis, sensory loss, and impairment of bowel and bladder function. In sheep and dog studies, intrathecal morphine infusions reliably produce catheter tip granulomas.2,3Within the human population, their development is less predictable. Morphine concentration and total daily dosage seem to influence the development of granulomas. The incidence increases with the duration of intrathecal therapy and has been reported as 0.4% after 2 yr of therapy and 1.16% after 6 yr of therapy.1The incidence of asymptomatic lesions may be much higher. In a cohort of seven patients, identification of a single symptomatic lesion led to imaging and identification of two additional asymptomatic lesions.4In a small canine study, combination therapy with morphine and clonidine was found to prevent granuloma development when the clonidine infusion was greater than 250 μg/day.1On this basis, some physicians who manage patients with intrathecal analgesic infusion systems recommend the routine use of clonidine as an adjunct with opioid to inhibit granuloma formation (personal observation, Kenneth Follett, M.D., Ph.D., Professor, Department of Neurosurgery, University of Iowa, Iowa City, Iowa, verbal communication, June 2004).
We describe a case in which continuous infusion of intrathecal clonidine did not prevent granuloma formation after subsequent addition of morphine to the intrathecal infusate. This observation suggests that clonidine may not have the same granuloma-inhibiting effect in humans that has been described in animal models.
After falling in 1995, a woman developed complex regional pain syndrome type I involving her left leg. Extensive conservative measures did not provide adequate relief of her pain, and consideration was given to intrathecal analgesic therapy. In April 1999, she underwent a successful trial (80–90% reduction of pain) with intrathecal clonidine, infused continuously over several days via temporary intrathecal catheter. Intrathecal opioid was not tested because of our previous experience that pain with a strong neuropathic component tends not to respond well to long-term to intrathecal opioid administration. After this successful trial, an intrathecal catheter and infusion pump were implanted in the patient in August 1999 (Medtronic model 8627-18, Medtronic catheter 8709 [89-cm catheter implanted]; Medtronic Neurologic, Minneapolis, MN). Initially, clonidine (1,000 μg/ml) was infused at 150 μg/day. This dose was selected based on the patient’s dose requirement during the trial. This rate was progressively increased, and after 6 months, she was receiving 950 μg clonidine per day. In March 2000, morphine was added to her infusion. Morphine (10 mg/ml) was infused at 1 mg/day because of inadequate pain relief with clonidine alone. Clonidine was continued, however, and the rate was reduced to 100 μg/day. During the next several months, the infusion rate of morphine was gradually increased to 10 mg/day with a parallel increase in clonidine. In June 2000, the infusion solution was changed to increase the concentration of morphine and clonidine to allow further titration and extend the period between pump refills. The new solution contained 17.5 mg/ml morphine and 1,300 μg/ml clonidine. After this change, the infusion rate was progressively increased until the patient reached a maximum dose of 15 mg morphine and 1,114 μg clonidine per day in April 2001 (table 1).
During the next year, the patient’s pain remained stable. Results of physical and neurologic examinations remained stable. In February 2002, she began having left lateral thigh pain. A lumbar radiculopathy was considered, and magnetic resonance imaging studies were obtained that demonstrated no significant findings. Physical examination revealed findings consistent with meralgia paresthetica. The patient was subsequently treated with a series of lateral femoral cutaneous nerve blocks, and her symptoms resolved. She continued with periodic pump refills until July 2003, when her left thigh pain returned. Again, she was diagnosed with meralgia paresthetica and treated with nerve blocks in July and September 2003. Her symptoms improved after each block.
In February 2004, the patient reported thigh pain during a pump refill visit and was referred to the Center for Pain Medicine and Regional Anesthesia for presumed recurrence of her meralgia paresthetica. In contrast to the pain she reported with her previous episodes of meralgia paresthetica, this pain was bilateral and was located more anteriorly on the thighs. She also reported bilateral leg weakness and increased pain with hip flexion. Physical examination at this time revealed no changes in her strength, sensation, or reflexes. However, given these new symptoms, lumbar and thoracic magnetic resonance imaging studies were completed.
These magnetic resonance imaging studies revealed a contrast-enhancing mass at the tip of the intrathecal catheter, consistent with intrathecal granuloma (figs. 1 and 2). The mass was located anterior to the spinal cord at the T11 level. It compressed the spinal cord and caused cord deformity. A neurosurgical evaluation was completed; given the absence of neurologic deficit and the anterior location of the lesion in relation to the cord (which would have made surgical resection of the mass difficult), conservative management was elected.
During the course of 15 days, the intrathecal morphine–clonidine dose was progressively reduced, and then the morphine–clonidine solution in the pump was replaced with preservative-free normal saline. Concurrent with the reduction in intrathecal analgesics, oral analgesic medications were increased. The lumbar and thoracic magnetic resonance imaging studies were repeated approximately 4 weeks after initiation of intrathecal saline infusion and revealed near complete resolution of the mass with restoration of the spinal cord to normal dimensions (figs. 3 and 4). At this time, the patient reported complete resolution of her leg weakness and leg pain. Retrospectively, the patient reported a history of intermittent urinary urgency and incontinence in the 6 months before discovery of the intrathecal mass that resolved entirely.
In most ways, this patient represents a typical granuloma patient. She had been seen frequently in the clinic for refilling of her infusion pump. She did not report significant change in her “regular” complex regional pain syndrome pain during her scheduled follow-up visits, her daily activities remained stable, and her neurologic examination results remained normal except for long-standing allodynia associated with her complex regional pain syndrome. The evaluation that revealed the intrathecal mass was prompted by new subjective changes in physical symptoms that were not accompanied by objective change in physical findings.
At the time the intrathecal mass was identified, the catheter had been present for 55 months; morphine had been infused for 48 months. This is somewhat longer than the average duration of therapy in patients described in the literature. In a case series of 41 patients, the median duration of therapy at the time of discovery was 24 months (range, 0.5–72 months).5Clonidine was the only component of the infusion for the first 7 months of therapy in the current patient. After morphine was added to the infusion, clonidine remained present continuously with the opioid infusion. The morphine was increased to a maximum of 15 mg/day (17.5 mg/ml) and continued for 35 months at this rate. In reviewing other case reports, this rate does not seem unusually large in either dose or concentration.1,4Sheep studies have shown the absence of granuloma formation at low morphine concentrations.3There is an apparent dose–response curve in dogs with higher doses and concentrations of morphine consistently producing granulomas. However, these studies are of short duration, and it remains unclear whether there is a safe (i.e. , non–granuloma-inciting) intrathecal dose of morphine for humans. Fully 30% of granulomas noted in the series of Yaksh et al. 2developed in patients receiving less than 10 mg morphine per day; 40% were using morphine concentrations less than 25 mg/ml. Limiting the infusion concentration and daily dose could potentially delay or prevent granuloma formation. Most patients require an increase in intrathecal dosing over time to maintain pain control.6If intrathecal opioids are limited, additional oral medications or the addition of adjunct nonopioid intrathecal agents will likely be required.
In dog studies, combination therapy with morphine and clonidine has been found to prevent granuloma development.1The protective effect of clonidine is dose dependent. At a morphine:clonidine ratio of 6:1 (1.5 mg morphine:250 μg clonidine), the incidence of granuloma development was the same as without clonidine. However, at a 3:1 ratio (1.5 mg morphine:500 μg clonidine), the incidence of granuloma formation is decreased. In this patient, clonidine was not protective against formation of a mass that is presumed to be an inflammatory granuloma. Clonidine was infused as a single agent before administration of opioid, and clonidine was administered continuously with infusion of opioid. Although the rate of clonidine infusion was variable, it was infused at the rate of 1,114 μg/day when the morphine infusion was at the maximum 15 mg/day. There are several possible explanations for the occurrence of a granuloma in this patient despite the prediction, based on animal models, that clonidine may protect against granuloma formation. The absolute dose of clonidine given to this patient was higher than that used in the dog studies, but the cerebrospinal fluid volume and flow in humans may be sufficiently different to require a much higher dose to be “protective.” The morphine:clonidine ratio in our patient was 13.5:1 (15 mg morphine:1,114 μg clonidine). To achieve the same “protective” 3:1 ratio noted in the canine studies and to continue a 15-mg/day morphine infusion, a 448% increase (5,000 μg) in the clonidine component would have been necessary. In absolute terms, this is far in excess of what is commonly used in clinical practice. It is possible that clonidine partially mitigated the granuloma-inciting effects of the opioid, given the fact that the patient’s granuloma was discovered after 48 months of opioid therapy, double the median duration of therapy at the time of discovery described in other case reports. In this case, the granuloma was not examined histologically. Clonidine may have affected the composition of the granuloma by alteration of inflammatory mediators or prevention of fibrosis. Additional studies are necessary to determine whether clonidine inhibits the formation of intrathecal opioid-induced granulomas and whether it might be a useful adjunct in intrathecal opioid therapy.
We elected, in consultation with the patient, to treat this mass nonsurgically. Our rationale for conservative management was based on the absence of frank neurologic deficit, the ventral location of the granuloma in the spinal canal (rendering surgical resection somewhat risky), the patient’s understanding of the situation and willingness to report any worsening of symptoms (with immediate return for follow-up evaluation, if necessary), and her reliability in monitoring her status and reporting changes to us. The granuloma was treated by ceasing intrathecal analgesic infusion, and within 6 weeks, the mass had resolved, with no residual sequelae. Surgical resection would have been undertaken if this patient had manifested clear-cut neurologic deficit, if symptoms had progressed despite cessation of intrathecal drug therapy, if radiographic improvement of the lesion did not occur after cessation of intrathecal drug therapy, or if we believed she would be unreliable or unable to report worsening of symptoms. Our patient’s good outcome should not be taken as approval for nonsurgical treatment of all granulomas. Neurologic deficit can develop and progress rapidly. Any patient with progressive neurologic symptoms or signs should be evaluated urgently or emergently for surgical treatment of the mass lesion.
Recommendations for detection, evaluation, and management of intrathecal opioid-induced granulomas have been published in a consensus statement.7This case supports the recommendations of the consensus panel, underscoring the need for a low threshold for evaluation and intervention in patients receiving intrathecal drug therapy with opioids who are suspected of harboring an intrathecal granuloma. It also highlights the importance of obtaining an interval history and examining the patient at the time of each pump refill, with special attention paid to eliciting symptoms of new pain, change in “usual” pain, or change in neurologic function. In addition, this case report demonstrates that the addition of clonidine to an intrathecal infusion of morphine does not obviate the need for clinical vigilance in regard to catheter tip inflammatory masses.