ObjectiveTo test the concept that externalized tunneled intrathecal catheters lead to a high risk of complications, such as meninigitis and epidural abscess, and therefore should not be used for durations of intrathecal pain treatment of > 1 week.DesignProspective, cohort, nonrandomized, consecutive, historical control trial.SettingTertiary care center, institutional practice, hospitalized and ambulatory care.PatientsTwo hundred adults (107 women, 93 men) with refactory cancer pain treated for 1–575 (median, 33; total, 14,485) days; 79 patients were treated at home for 2–226 (median, 36; total, 4, 711) days. All patients had died by the close of the study.InterventionsInsertion of intrathecal tunneled nylon (Portex) catheters (223 in 200 patients) with Millipore filters. The catheter hubs were securely fixed to the skin with steel sutures. Standardized care after insertion: (a) daily phone contact with the patients, their families, or the nurses in charge; (b) weekly dressing change at the tunnel outlet by the nurses; (c) refilling of the infusion containers by the nurses; (d) exchange of the infusion systems when empty (within 1 month) and of the antibacterial filter once a month by specially instructed Pain Department nurses. All contact between the connections of the syringes, cassettes, and needles with the operator's hands was carefully avoided during filling and refilling of the infusion containers and exchange of the antibacterial filters; no other aseptic precautions were taken.Main Outcome MeasuresWe recorded the rates of perfect function and complications of the systems. The rates of complications recorded in this study with externalized tunneled intrathecal catheters are discussed and compared with the rates reported in the literature with externalized (tunneled and nontunneled) epidural and intrathecal catheters, as well as with internalized (both epidural and intrathecal catheters connected to subcutaneous ports, reservoirs, and pumps.ResultsThe following rates (as a percentage of number of patients) of perfect function and complications of the systems were recorded (the ranges of rates reported in the literature are given in parentheses): perfect function of the system, 93% (31–90%); accidental injury of an unknown epidural tumor followed by an epidural hematoma, 0.5% (04%); skin breakdown at the insertion site, 2% (2–50%); postdural puncture headache, 15.5% (10%); external leakage of CSF, 3.5% (427%); CSF hygroma (“pseudomeningocele”), 1.5% (4− 6.25%); hearing loss and MBniBre-like syndrome, 0% (12%); pain on injection, 0% with continuous infusion and 4.5% with intermittent injections (3–36% with intermittent injections); catheter tip dislodgement, 1.5% (633%); catheter (system) occlusion, 1% (3–12%); accidental catheter withdrawal, 4% (3–22%); catheter (system) leakage, 1.5% (2.1–26.6%); all mechanical complications, 8.5% (1044%); local (catheter entry site) infection, 0.5% (2–33%); catheter track infection, 0% (625%); epidural abscess, 0% (0.625%); meningitis, 0.5% (1–25%); systemic infection, 0% (3%); incidence of all infections (nhreatment days), 1 /7,242 ( 1 / 168− 1 /2,446).ConclusionsIn our population and with the technique of insertion and care reported here, the use of externalized tunneled intrathecal catheters has not been associated with higher rates of complications when compared with earlier reported rates of externalized epidural catheters and internalized (both epidu- ral and intrathecal) catheters connected to subcutaneously implanted ports, reservoirs, and pumps. The opinion that the use of externalized tunneled in- trathecal catheters should be restricted only to patients who need pain treat- ment for