Abstract
INTRODUCTION: Intrathecal drug delivery systems (IDDS) represent an advanced option for refractory craniofacial cancer pain, though optimal catheter tip positioning remains debated between anterior cisterna (AC) and high cervical (HC) approaches. METHODS: This multicenter retrospective cohort analyzed 108 patients with severe craniofacial cancer pain undergoing IDDS catheterization (33 AC; 75 HC) between January 2018 and December 2024. Propensity score matching (1:2 ratio) created balanced cohorts (33 AC versus 66 HC). Primary outcome was pain reduction (numerical rating scale), with minimum clinically important difference defined as ≥ 1.5-point reduction. Secondary outcomes included opioid requirements, intrathecal morphine dosing, conversion ratios, and safety. RESULTS: After matching, both approaches demonstrated comparable pain reduction, with 90.9% of AC and 92.4% of HC patients achieving minimum clinically important difference at 3 months (risk difference: 1.5%, 95% confidence interval [CI] -9.8 to 12.8%). HC patients required higher intrathecal morphine doses during the first 2 weeks (week 1: 285 versus 220 mcg/day; ratio: 0.77, 95% CI 0.65-0.91), with significantly different oral-to-intrathecal conversion ratios (week 1: 886 versus 667; ratio: 1.33, 95% CI 1.12-1.58); these differences resolved by 1 month. Complete systemic opioid discontinuation at 3 months was comparable between groups (63.6% AC versus 59.1% HC). Safety profiles differed, with AC patients experiencing more procedure-related complications (postdural puncture headache: 18.2% versus 4.5%), while HC patients showed trends toward more opioid-related adverse events. CONCLUSIONS: Both AC and HC catheterization provide comparable pain control for refractory craniofacial cancer pain. The AC approach offers superior initial delivery efficiency but requires advanced technical skill, while the HC approach provides comparable long-term outcomes with a more familiar technique. Selection should be guided by institutional expertise and patient-specific risk profiles.