Efficacy and Technical Complications of Long-Term Continuous Intraspinal Infusions of Opioid and/or Bupivacaine in Refractory Nonmalignant Pain: A Comparison Between the Epidural and the Intrathecal Approach with Externalized or Implanted Catheters and Infusion Pumps

Objective: To compare efficacies, failure rates, and technical complication rates of intraspinal treatments in patients with "refractory" nonmalignant pain conditions in relation to the approach (epidural/intrathecal), the drug (opioid/opioid-bupivacaine or bupivacaine), and the type of system used (externalized/internalized). In these comparisons, recent data from a companion paper (Nitescu et al., Clin J Pain 1998;14:17-28) were used as a reference to be compared with data from a literature review of different intraspinal treatment modalities in nonmalignant pain. Design: Prospective, cohort, nonrandomized, consecutive trial. Setting: Tertiary care center, institutional practice, hospitalized, and ambulatory care. Patients: Five groups according to treatment modality: (a) externalized, long-term intrathecal nylon catheters, connected to external, electronic infusion pumps (companion paper), n = 90; (b) internalized, long-term intrathecal catheters (Silastic®) connected to implanted SynchroMed® pumps, n = 330 (literature review); (c) externalized, "short-term" epidural catheters for "temporary" infusions, n = 565 (literature review); (d) externalized, long-term epidural catheters, n = 50 (literature review); (e) internalized, long-term epidural catheters, n = 111, connected to implanted systems: Port-A-Cath® injection ports, n = 58; Infusaid® pumps, n = 46; and SynchroMed® pumps, n = 7 (literature review). Interventions: In reviewing the literature, we found 21 studies that reported on the intraspinal (epidural or intrathecal) administration of opioids with or without local anesthetics (usually bupivacaine). These studies were analyzed with respect to the rates of the variables satisfactory pain relief (efficacy), failures, and technical complications. A rate is the number of observations of a variable divided by the number of patients or the number of catheters or infusion systems, as logically indicated (e.g., the numbers of complications, such as epidural abscess and meningitis, were related to the number of patients and those of catheter occlusion or leakage to the number of the catheters). The variables were expressed as the means of the rates of a variable from studies belonging to various treatment modalities: approach (epidural vs. intrathecal), duration (short vs. long term), drugs administered intraspinally (opioid vs. opioid and/or local anesthetic), and type of infusion system (externalized vs. internalized). Further, the sums of all observations of one variable in different studies with various treatment modalities were related to the corresponding sums of the patients (alternatively, catheters or implanted devices). The proportions of these sums were tested for significance in relation to treatment modality. Main Outcome Measures: Comparative rates of successful intraspinal treatment and its failures and complications. Results: (a) The intrathecal approach, compared with the epidural approach, was associated with higher rates of satisfactory pain relief for both externalized (86/90, 95% vs. 17/40, 42.5%, p < .0001) and internalized (295/336, 89% vs. 33/56, 59%, p < .0001) catheters; higher rates of treatment failures with externalized epidural catheters than with internalized intrathecal catheters (24/47, 51%, vs. 36/338, 11%, p < .0001); lower rates of treatment failures with internalized intrathecal catheters than with internalized epidural catheters (36/338, 11% vs. 29/76, 38%, p < .0001); higher rates of system replacement with internalized epidural catheters than with internalized intrathecal catheters (23/32, 72% vs. 6/49, 12%, p < .0001); higher rates of system removal with internalized epidural catheters than with internalized intrathecal catheters (22/49, 45% vs. 5/49, 10%, p < .001); higher rates of catheter-related complications with epidural than with intrathecal catheters (dislodgement 13/126, ∼10% vs. 6/150, 4%, p < .05; leakage 5/51, ∼10% vs. 1/116, 0.9%, p < .05; obstruction 21/75, 28% vs. 1/101, 1%, p < .0001). (b) The rates of satisfactory pain relief were lower in the patients treated with epidural opioids (43/89, 48%) than in those treated with intrathecal opioids (285/323, 88%, p < .0001), intrathecal opioids and bupivacaine (96/103, 93%, p < .0001), and epidural opioids and bupivacaine (7/7, 100%, p < .05). (c) Significantly higher rates of catheter dislodgement (8/124, ∼6% vs. 1/152, ∼0.7%, p < .0001) and obstruction (21/75, 28% vs. 1/101, ∼1%, p < .0001) and lower rates of accidental catheter withdrawal (0/441, 0% vs. 11/152, ∼7%, p < .0001) were found with internalized than with externalized infusion systems. (d) Use of implanted SynchroMed® pumps connected to implanted intrathecal Silastic® catheters showed high rates of catheter (7-44%) and pump (6-44%) failures and no advantages over the externalized intrathecal catheters connected to electronic infusion pumps with regard to rates of pain relief, treatment failure, and system-related complications. Conclusions: (a) The intrathecal approach appeared to be superior to the epidural approach regarding satisfactory pain relief, treatment failure, and complication rates. (b) Administration of epidural opioids was associated with lower rates of satisfactory pain relief than were intrathecal opioids or opioids and bupivacaine, and epidural opioids and bupivacaine. (c) The use of totally implantable Synchromed® pumps seemed to be associated with high reported rates of technical failures.