Intrathecal narcotic infusion pumps for intractable pain of chronic pancreatitis: a pilot series.

OBJECTIVES

The aim of this study was to evaluate the efficacy of intrathecal narcotics pump (ITNP) as an alternative treatment for patients with pain from chronic pancreatitis (CP). ITNP offers the advantages of reversibility, lower total narcotic dose, and the pancreas remaining intact.

METHODS

Thirteen patients (8 female, 5 male), with mean age 40.6 years (s.d. 9.6 years), who had experienced intractable upper abdominal pain from CP were reviewed. Each patient had multiple other failed treatment modalities, including partial pancreatic resection (n = 6). They were offered ITNP after a successful intraspinal opioid trial. Etiologies of CP included idiopathy (n = 3), cystic fibrosis (n = 2), alcohol (n = 2), and pancreas divisum (n = 6).

RESULTS

The median duration of severe, intractable pain prior to ITNP was 6 years (2-22 years). The median follow-up time after ITNP was 29 months (range, 7-94 months). The ITNP was in situ for a mean duration of 29 months (range, 0.5-94 months). Seven patients had pump exchange or removal for various reasons; improvement of pain at month 53 (n = 1), meningitis (n = 1), meningitis with subsequent replacement (n = 1), pump failure at month 31, 68, 79, and 84 (n = 4). There were no deaths. The mean pain score prior to implantation (score = 8.3, s.d. = 0.9) was significantly higher than 1 year after (score = 2.7, s.d. = 1.9) (P < 0.01) and last follow-up (score = 0.75, s.d. = 2.1) (P < 0.01). The median oral narcotic dose before and 1 year after ITNP were morphine sulfate equivalents 337.5 mg per day (range, 67.5-1,320) and 40 mg per day (range, 0-1,680), respectively (P < 0.01). Two patients were considered failures, as they still require a high dosage of both oral and intrathecal medications to control their pain, despite significant pain-score improvement. One patient who was excluded due to meningitis was also considered a failure. Therefore, the overall success rate of ITNP based on an intention-to-treat analysis was 76.9% (10/13). The major complications of ITNP were central nervous system infection requiring pump removal (n = 1), cerebrospinal fluid leak requiring laminectomy (n = 1), and perispinal abscess with bacterial meningitis requiring pump removal (n = 1).

CONCLUSIONS

This study shows the many risks and benefits of ITNP. A longer follow-up is awaited; such pumps appear to be one alternative to aggressive surgical intervention. Failed ITNP trials leave other options open. Therapeutic trials directly comparing pancreatectomy, ITNP, and implanted nerve stimulators are of interest.