Skip navigation
Acute Pain | Chronic Pain | General

Neurolytic coeliac plexus block (NCPB) for cancer pain

Clinical bottom line:

Clinical bottom line: based on data from uncontrolled trials, neurolytic coeliac plexus block is an effective method for relieving pain associated with pancreatic and non-pancreatic intra-abdominal malignancies, with approximately 90% of patients achieving at least partial pain relief short term (two weeks) and long term (three months and beyond). There appears to be no added benefit from NCPB with any form of radiological guidance.

Common transient adverse effects were local pain, diarrhoea and hypotension. More severe adverse effects included neurological and non-neurological adverse effects, both of which occurred in 1% of patients.

Between 2% and 5% of patients with advanced cancer receiving hospice care have nerve blocks as part of cancer pain management. Neurolytic blocks, i.e. those which destroy the nerve, are only used in patients with short life expectancies or where other simpler techniques have failed. In this situation neurolytic coeliac plexus block (NCPB) is used to treat pain associated with upper abdominal cancer.

Systematic review

Eisenberg E, Carr DB, Chalmers TC. Neurolytic coeliac plexus block for treatment of cancer pain: a meta-analysis. Anesth Analg. 1995; 80: 290-5.

Inclusion criteria were randomised and non-randomised trials of NCPB for cancer pain; controlled and uncontrolled trials; full journal publication; English language.

Reviewers analysed the results of randomised controlled trials separately from those of other studies. Pain and other data were extracted. Pain data were categorised as either complete pain relief, partial pain relief and minimal/no relief. Results were categorised as acute (up to two weeks) or long-term (two or more weeks). Where pooling was possible, the estimate of the success rate of the intervention was assessed by calculating weighted means using a random effects model.


Of the included trials which specified cancer type, 707/1117 patients had pancreatic cancer, and 410/1117 had non-pancreatic intra-abdominal malignancies. Duration of patient pain was approximately two to seven months, and where reported (four trials) was classified as severe.

Randomised controlled trials

Two trials were included. One compared three different types of NCPB technique and the second compared NCPB with oral analgesic. Unfortunately reviewers did not fully report the findings of these trials. In one trial 10/10 patients report at least partial pain relief at 2 weeks and 7/10 at three to 10 weeks (control information not given). In the second trial 70% to 80% of patients reported at least partial pain relief at two weeks, and 60% to 75% at three months and beyond (control information not given).

Uncontrolled trials

Short-term analgesic efficacy: 18 trials in 989 patients were included. 89% of these had at least partial relief. Of the patients experiencing at least partial relief, 59% experienced complete relief by two weeks.

Long-term efficacy: 90% of 273 patients experienced at least partial relief at three months or beyond.

Six trials with 53 patients reported pain at time of death. 73% to 92% of patients had at least partial relief.

There were no apparent differences in pain relief in pancreatic compared with non-pancreatic cancer.

Procedure methods

Most patients underwent radiologically guided NCPB (including computed tomography, radiography and less commonly fluoroscopy or ultrasound). Overall, the success rates (at least partial pain relief) varied from 86% to 95%.

NCPB with no radiological guidance was reported for 238 patients from uncontrolled trials. There was a mean success rate of 95% (92 to 98 confidence interval).

One randomised controlled trial reported a 90% success rate with radiographic guidance. A further 271 patients from uncontrolled trials had a mean success rate of 91% (87 to 94 confidence interval).

Computed tomography was reported for 271 patients from uncontrolled trials. There was a mean success rate of 88% (83 to 97).

One randomised trial reported a 70% to 80% success rate with fluoroscopy. A further 36 patients from an uncontrolled trial had a mean success rate of 86%.

Adverse effects

There were three commonly reported transient adverse effects: local pain was reported in 96% of patients (two trials) diarrhoea in 44% of patients (five trials) and hypotension in 38% of patients (ten trials).

More severe effects were reported in 13 trials. Neurological complications such as lower extremity weakness and paresthesia, epidural anaesthesia and lumbar puncture were reported in 1% of patients. Significant non-neurological adverse effects, including pneumothorax, shoulder, chest and pleuritic pain, hiccuping and hematuria occurred in a further 1% of patients.

It was not possible to establish whether different procedures were associated with particular adverse effects.

Related topics