Intrathecal Pump Selection for Failed Back Surgery Syndrome: Opioid, Ziconotide, or Both
A man whose back pain never resolved after three surgeries is being considered for an implanted pump that delivers medication directly to his spinal fluid, and the device itself is the easier decision — what actually goes in it is not.
W.H., a 52-year-old man, worked as a commercial roofer for over two decades before a fall from a ladder six years ago fractured two lumbar vertebrae, the first of what became three spine surgeries — decompression, then fusion, then a revision fusion eighteen months ago after hardware failure. None of the three surgeries resolved his pain, a pattern his surgeon and pain specialist now both formally describe as failed back surgery syndrome. He has tried and exhausted the standard systemic ladder: gabapentin, duloxetine, high-dose systemic opioids that provided meaningful but incomplete and increasingly side-effect-laden relief, a spinal cord stimulator trial that helped his leg pain but did little for the axial back pain that dominates his day, and multiple epidural steroid injections with only brief benefit each time. He also carries a history of major depressive disorder, diagnosed a decade before his first surgery, well controlled for several years on sertraline with no hospitalizations and no psychotic features ever documented. He and his pain specialist are now discussing an intrathecal drug delivery system — an implanted pump delivering medication directly into his spinal fluid — as the next real option, deliberately distinct from the cancer-pain population his site's existing pain-service coverage already addresses elsewhere.
The pump itself is, at this point, the less contested part of the decision; what actually goes into it carries the real disagreement. Intrathecal morphine remains the most extensively used agent, with decades of accumulated real-world experience across thousands of patients, and his own partial (if imperfect) response to systemic opioids offers some signal that an opioid-responsive component genuinely exists in his pain. Ziconotide, a non-opioid N-type calcium channel blocker delivered the same intrathecal route, carries real randomized-trial support of its own — Rauck and colleagues' pivotal trial established significant pain reduction against placebo — and the current Polyanalgesic Consensus Conference algorithm lists it alongside morphine at Line 1A, with a separate consensus point favoring ziconotide as first-line in chronic non-cancer pain unless contraindicated — a preference built on its freedom from intrathecal opioid therapy's tolerance and respiratory-depression risk. That conditional clause is where W.H. stops being a straightforward fit. What that same PACC algorithm does not let get glossed over, though, is ziconotide's own labeled risk: neuropsychiatric adverse effects, including psychosis, in a patient whose psychiatric history — however well controlled for years — is real and on the chart.
Intrathecal therapy planning conference
Start with intrathecal morphine as the trial agent. It's the most extensively used intrathecal drug we have, with decades of real-world data behind it, and his partial systemic opioid response — imperfect, but real — is a signal that an opioid-responsive component exists in his pain that we shouldn't ignore just because the systemic route hasn't worked well enough.
I'd start with ziconotide instead. Rauck's pivotal randomized trial established real efficacy against placebo, it carries no risk of tolerance or respiratory depression the way intrathecal opioids do over time, and the PACC algorithm carries a specific consensus point favoring ziconotide as first-line in chronic non-cancer pain unless contraindicated — that's not my own ranking, it's where the field's own consensus has moved. Both drugs sit at Line 1A, so this is a preference within a shared tier, not morphine being demoted out of it.
Decades of experience with morphine is real, but "most familiar" and "guideline-preferred first-line" aren't the same claim, and the field moved for a specific, named reason — tolerance and respiratory risk — that his own history doesn't erase.
Both of your arguments are real, and I don't think either of you has weighed the thing that actually changes this for him specifically: ziconotide carries a labeled risk of neuropsychiatric adverse effects, including psychosis. He has a real psychiatric history — well controlled for years, but real, on his chart, not a footnote. That risk needs to sit directly in this decision, not get treated as background noise next to an efficacy argument.
The PACC algorithm itself weighs psychiatric history as a named factor, not an afterthought — citing ziconotide's guideline preference without also citing that same guideline's own caution in a patient like him is citing half the algorithm.
Agreed: intrathecal morphine as the initial trial injection, given his real partial systemic opioid response and his psychiatric history weighing against ziconotide as the first agent specifically for him; ziconotide retained as the next option if morphine's trial doesn't produce adequate relief, with a formal psychiatric baseline and monitoring plan established in advance of that decision rather than after. The psychiatrist's individualized risk argument was what shifted the pharmacologist's opening preference, not a rejection of ziconotide's guideline standing generally.
Not agreed: whether the PACC algorithm's current first-line preference for ziconotide should apply more strongly to future patients without his specific psychiatric history, or whether this case reveals the guideline itself under-weights psychiatric screening as a formal gate rather than a soft consideration. Logged as a question for the pain service's broader intrathecal-therapy protocol, not resolved for this one patient's plan.