EXTRAMEDULLARY intrathecal granuloma formation is a rare complication of morphine administration via  implanted drug delivery systems.1–5We present a unique case of a granuloma adherent to the conus medullaris presenting as cauda equina syndrome. Practitioners should not assume that implantation of intrathecal catheters with their tips located at or below the conus medullaris will eliminate all risk of neurologic sequelae associated with inflammatory granulomas.

Over the course of 15 yr, a 47-yr-old male with a history of traumatic L1 compression fracture underwent five lumbar and thoracic surgeries with instrumentation. He subsequently underwent intrathecal morphine pump (SynchroMed®, Medtronic, Minneapolis, MN) placement in 1999 to control his pain. He presented after 5 yr of adequate pain relief at an intrathecal dose of 25 mg of morphine per day (stable over the previous 12–18 months). Beginning 4 months before admission, the patient noted new onset and increasing numbness and sharp burning pain in his left lower extremity. During this time, his physician had liberally increased the patient’s oral opioid administration and administered neuropathic pain medication, but no improvement of his symptoms occurred. The patient was admitted with severe unremitting neuropathic pain in his left lower extremity. He was afebrile with a normal leukocyte count. Neurologic examination revealed hypesthesia (without allodynia) from L2-S1 in the left lower extremity. Motor strength and reflexes were preserved bilaterally. The patient had no complaints of bowel or bladder incontinence or sexual dysfunction. Lumbar spine magnetic resonance imaging revealed a T2 hypointense, T1 hyperintense, intrathecal, extramedullary, ovoid 7-mm mass to the left and lateral to the conus medullaris with displacement of the conus to the right. The conus itself demonstrated a central abnormal T2 hyperintensity signal extending craniocaudally for one vertebral level, suggesting cord edema (fig. 1, A and B). This lesion was consistent with a catheter tip granuloma.

After surgical consent, the patient was taken to the operating room for exploration and removal of the granuloma. The catheter entered the dura at approximately L3 with a cranial course to L1. A midline durotomy at the level of L1 revealed the catheter tip to be located within the cauda equina along the left side (fig. 1C). A brownish extramedullary mass adherent to the conus medullaris surrounded the catheter at its proximal port (fig. 1D). Several of the nerve roots of the cauda equina were adherent to the granuloma (fig. 1D) and were carefully dissected away. On further inspection, the center of the mass contained a thick, purulent-appearing substance that was sent for culture. Solid components of the necrotic mass were plugging the more distal catheter tip openings, requiring blunt dissection for removal of the tip. The necrotic mass was removed via  sharp dissection (fig. 1, C, D, E). Microscopic examination of the lesion revealed a chronic inflammatory process with necrotic neural tissue. Because bacterial contamination of the entire catheter and pump could not be ruled out, both were removed.

The patient was placed on broad-spectrum antibiotics because of the presence of purulent-appearing material within the center of the granuloma. Consultation with infectious disease specialists suggested the need for repetitive lumbar puncture to determine the presence of occult cerebrospinal fluid (CSF) infection before any consideration of reimplantation, although signs and symptoms of infection never occurred. None of the cultures from the granuloma, catheter tip, pump pocket, or fluid from the pump reservoir revealed bacterial growth.

Postoperatively, the patient’s pain was very difficult to manage (average hourly verbal numerical pain score was 9 on a 1–10 scale) despite using 450–550 mg of hydromorphone on a daily basis via  intravenous patient-controlled analgesia. Because of poor pain control despite access to these extraordinary amounts of intravenous opioids, consideration was given to reservoir implant for intracerebroventricular administration of opioids.6This was abandoned because of the substantive risk of infection.6Despite the patient’s increased risk for recurrent granuloma, an intrathecal drug delivery system was surgically placed such that the catheter pierced the dura at L3 with a caudal course to the level of L5. The patient recovered well from this procedure and was weaned off intravenous opioids without withdrawal. All neurologic symptoms and signs resolved completely, and the patient was discharged home receiving 12.5 mg of intrathecal morphine per day.

Several weeks after discharge, the patient developed new and sudden onset left S1 radicular pain. Repeat magnetic resonance imaging demonstrated that the catheter tip had migrated into the left lateral recess at S1 (fig. 2). With the patient awake, the catheter was repositioned by percutaneous withdrawal under fluoroscopy. The left S1 radicular pain immediately and completely resolved.

Before the identification of catheter-associated granulomas, it had been customary to place catheter tips at thoracic levels. A shift in practice to catheter placement within the lumbar cistern has now occurred in an attempt to eliminate the risk of neurologic sequelae associated with the development of granulomas. The significance of this case report is that neurologic dysfunction may still occur with catheter tip placement at or below the level of the conus medullaris and the presentation described in this case report will likely become more common, given the shift in practice to lumbar catheter tip placement.

Despite the atypical location of the granuloma, the patient had a typical presentation for intrathecal catheter-associated granuloma: sensory changes and increasing pain that was refractory to escalating doses of analgesics. Sensory abnormalities and increased pain are very common presentations of granulomas that can occur with granulomas affecting any part of the spinal neuraxis.3Vigilance, with strict attention to symptoms (e.g. , increasing pain or new pain) and neurologic signs, is still warranted despite lumbar catheter tip placement to allow early detection of granulomas. Motor deficits typically occur later in the natural history of the progression of granulomas after prodromal symptoms have been overlooked.1,3Motor dysfunction may still occur with lumbar catheter tip placement.

Purely anatomic considerations would suggest that catheter placement below the conus medullaris could prevent paresis or paralysis from cord compression. However, motor dysfunction is still possible. Typically, lesions of the conus medullaris (in this case, compression) spare motor and reflex function in the lower extremities but are associated with dense perineal anesthesia (S3-S5) and prominent bowel, bladder, and sexual dysfunction. Lesions of the cauda equina are characterized by neurologic signs and symptoms referable to the involved nerve roots, manifesting as radicular pain, asymmetric lower extremity hypesthesia, anesthesia, motor dysfunction, and variable areflexia. There is relative preservation of bowel and bladder function. Mass lesions in the lower spinal canal can produce a clinical presentation that is a mixture of conus medullaris and cauda equina syndromes. The patient in this report did not present with conus medullaris syndrome despite compression and adherence of the granuloma to the conus. The patient did present with cauda equina syndrome without a motor deficit, although motor abnormalities are possible.

Current hypotheses as to the etiology of catheter granulomas involve the combined contributions of high drug concentration, duration of high concentration infusion, catheter tip location, and poor CSF flow patterns.1–3Indeed, this patient had been receiving a high concentration of morphine for at least 2 to 3 yr. In a review of 41 patients with intrathecal granulomas, Coffey and Burchiel reported that 39% of the patients received ≥10 mg per day and another 39% of patients received ≥25 mg per day.1Magnetic resonance imaging models of pulsatile CSF flow predict the ventral cervical, dorsal thoracic, and ventral lumbar spinal canal to have maximal CSF flow velocities.7,8Such areas of high flow would promote good mixing of drug with CSF. Although there is a larger area of CSF in juxtaposition to neural tissue within the lumbar canal in comparison with the thoracic spinal canal, regions of low flow, pooling, and poor mixing of drug do exist.1 

It is important to comment on the rationale for surgical intervention in this patient. Once a granuloma has been identified, the physician must decide whether to remove the granuloma and all or part of the drug infusion system. Treatment options include cessation of the infusion with shrinkage or disappearance of the granulomas within 2–5 months,1,3percutaneous withdrawal of the catheter, and partial or complete surgical removal of the granuloma and infusion system.

Discontinuation of drug infusion is usually reserved for patients with minimal signs and symptoms and small granulomas. In the present situation, the patient was in extremely severe pain that was refractory to neuropathic pain medications and opioids. Cessation of drug administration would have mandated an intrathecal “drug holiday.” Over the 12–18 months before surgery, the patient had received a stable dose of 25 mg of morphine per day. It was anticipated that the patient’s pain would be exceedingly difficult to assuage without intrathecal drug administration, and in retrospect, this was true.

In cases where motor dysfunction is not present, percutaneous revision, repositioning or replacement of the catheter have been advocated.3Such procedures are to be performed in awake patients with fluoroscopic guidance, continuous lower extremity monitoring of motor and sensory function, the utilization of only gentle traction, and abandonment of the procedure if “force beyond gentle traction is required and/or if the patient reports intraoperative symptoms that indicate displacement of the mass.”3Percutaneous withdrawal of the catheter was considered but not pursued during the patient’s initial presentation because of the exceedingly severe and unremitting neuropathic pain.

In retrospect, percutaneous catheter withdrawal might have initially failed in this patient because of the extension of the granuloma into the catheter’s fenestrations, although this is speculation. Similarly, because of its anatomic location, simply pulling the catheter back without durotomy and inspection might have resulted in neurologic deficit resulting from adhesions between the catheter, granuloma, conus medullaris, and several of the nerve roots of the cauda equina. However, this is speculation also. As was demonstrated subsequently with resolution of the S1 radicular pain, percutaneous catheter withdrawal can be effective and safe.

In summary, we present a case of catheter-associated granuloma adherent to the conus medullaris presenting with cauda equina syndrome. Despite the shift in current practice away from thoracic catheter tip placement and its hypothesized benefits on an anatomic basis, this case study demonstrates two complications arising from catheter tip placement within the lumbar cistern. Practitioners must not assume that lumbar catheter tip placement for intrathecal drug infusion is devoid of neurologic sequelae. Regular attention to symptoms and neurologic signs is still warranted despite lumbar catheter tip placement to allow early detection of granulomas.

Coffey RJ, Burchiel K: Inflammatory mass lesions associated with intrathecal drug infusion catheters: Report and observations on 41 patients. Neurosurgery 2002; 50:78–87
Yaksh TL, Hassenbusch S, Burchiel K, Hildebrand KR, Page LM, Coffey RJ: Inflammatory masses associated with intrathecal drug infusion: A review of preclinical evidence and human data. Pain Med 2002; 3:300–12
Hassenbusch S, Burchiel K, Coffey RJ, Cousins MJ, Deer T, Hahn MB, DuPen S, Follett KA, Krames E, Rogers JN, Sagher O, Staats PS, Wallace M, Willis KD: Management of intrathecal catheter-tip inflammatory masses: A consensus statement. Pain Med 2002; 3:313–23
McMillan MR, Doud T, Nugent W: Catheter-associated masses in patients receiving intrathecal analgesic therapy. Anesth Analg 2003; 96:186–90
Follett KA, Naumann CP: A prospective study of catheter-related complications of intrathecal drug delivery systems. J Pain Symptom Manage 2000; 19:209–15
Lazorthes YR, Sallerin BA-M, Verdie JCP: Intracerebroventricular administration of morphine for control of irreducible cancer pain. Neurosurgery 1995; 37:422–9
Grietz D, Franck A, Nordell B: On the pulsatile nature of intracranial and spinal CSF circulation demonstrated by MR imaging. Acta Radiol 1993; 34:312–28
Thomsen C, Stahlberg F, Stubgaard M, Nordell B: Fourier analysis of cerebrospinal fluid flow velocities: MR imaging study. The Scandinavian Flow Group. Radiology 1990; 177:659–65