BACKGROUND: Intrathecal (IT) drug therapy with implanted pumps is an effective treatment modality for chronic pain and/or spasticity, especially after non-invasive treatment has failed. Long-term use of intrathecal opioids may cause formation of inflammatory masses at the tip of intrathecal catheters, possibly leading to neurological deficits and/or catheter revision. OBJECTIVE: We aimed to identify risk factors for catheter-tip granuloma (CG) formation. STUDY DESIGN: Retrospective study. SETTING: Tertiary Spine Centers in Germany and SwitzerlandMETHODSWe retrospectively reviewed data at 2 Swiss centers (Kantonsspital St. Gallen, Swiss Paraplegic Centre Nottwil) between 01/1994 and 10/2013. Collected data were age at operation, gender, smoking status, previous spinal operations, spinal level of catheter-tip, clinical symptoms, catheter testing with contrast agent, applied drugs, drug concentration, as well as cumulative daily drug dosages. RESULTS: Thirteen patients with a mean age of 52.6 years and CG formation after a mean of 6.9 years of follow-up were identified and compared to 54 patients of similar age and length of follow-up (48.6 years, P = 0.535; follow-up 5.3 years, P = 0.236) without CG. In the analysis of risk factors, catheter ending in the middle thoracic spine (Th4-8; 38.5 vs. 6.5%; P = 0.010), previous spinal surgery (75 vs. 41%; P = 0.051), and chronic pain as an underlying primary symptom for IT drug therapy (100 vs. 56%, P = 0.003) were associated with CG formation. IT drug therapy for spasticity appeared to be much less associated with CG formation (0 vs. 44%, P = .0003). As the symptomatology is closely related to the medical treatment applied, patients with CG were more likely to be treated with IT morphine (77 vs. 20%; P < 0.001), and as tendency with IT clonidine (54 vs. 26%; P = 0.092) and IT bupivacaine (46 vs. 20%; P = 0.077). Average in-pump morphine concentration (30.3 vs. 19.5 mg/mL; P = 0.05) as well as average daily dose of morphine (12.5 vs. 6.2 mg/d; P = 0.037) were significantly higher in the CG group. Smoking could not be identified as risk factor for CG formation. LIMITATIONS: Limitations include the retrospective approach, the limited group size of granuloma patients, as well as missing data in the investigated patient groups. CONCLUSION: Our patient cohort with CG differed in some features, of which some like catheter localization, choice, dosage, and the concentration of drugs are potentially modifiable. These results could contribute to the prevention of CG in the future.