Interventional pain management includes a range of specialised invasive techniques for treating pain that are usually offered by either anaesthetists or interventional radiologists. The most common techniques used for palliative care problems include spinal analgesia and nerve blocks. The availability of interventional pain management varies according to the setting of care and the local availability of skilled and experienced providers. The place of these techniques within a hierarchy of pain management options is still evolving.
What is known
Spinal and intrathecal routes of analgesia are sometimes used for difficult pain management problems. However the widespread acceptance of intraspinal analgesia  is based mainly on preclinical data and expert opinion. A survey and systematic review of the literature revealed considerable variation in practice, scant evidence to support clinical decision making, and little consensus on the medications to be used. [2,3] These reviews and a recent update  provide little guidance about the indications for choosing the intraspinal route for analgesia in palliative care patients. However, in another systematic review,  the two main indications consistently identified were intractable pain not controlled by other conventional medical routes and / or side effects from conventional pain management strategies preventing dose escalation. Intraspinal analgesia appears to be as effective as conventional medical management and may have fewer side effects.
A systematic review compared epidural, subararachnoid and intracereboventricular (ICV) routes of administration for spinal analgesia and found that the ICV route was at least as effective and associated with less adverse effects. However all the data are from uncontrolled trials. 
Catheter associated problems may occur with longer-term use of intraspinal opioids,  however a systematic review looking specifically at this issue has concluded that these events are rare in both hospital and community settings. 
The best evidence is for spinal analgesia using morphine as a first line drug. There is also limited evidence to support use of morphine / bupivicaine if the pain has a neuropathic component, or of morphine / clonidine, and hydromorphone.  If pain is not responsive to opioids, consensus guidelines support the use of ziconotide and clonidine.  For other medications and combinations the evidence is not strong, and further research is required.
Veterbroplasty / kyphoplasty
Vertebroplasty and kyphoplasty both involve the injection of cement into vertebrae affected by compression fractures with the aim of improving pain. The use of vertebroplasty as a technique to treat malignant fractures is controversial.  Despite being associated with improvements in pain and disability,  the literature on these interventions is not robust, and in particular there is a lack of RCTs or prospective randomised data to explore the contribution of the placebo effect. One of these systematic reviews raises concerns about the rate of severe complications from vertebroplasties, which have been directly associated with a number of fatalities.  It is suggested that such interventions may carry increased risks in patients with advanced malignancy compared with those who have osteoporosis, and that this may perhaps be related to the volume of cement injected. The authors argue that, based on this data, vertebroplasty or kyphoplasty should not be thought of as minimally invasive techniques in this population.
Sympathectomy is a procedure that damages part of the nervous system (the sympathetic chain) in order to improve neuropathic pain or complex regional pain syndrome. A recent systematic review has found almost no evidence to support the procedure – one small study of reasonable quality was included and no improvement could be demonstrated after sympathectomy, but there is a significant risk of adverse effects. 
Coeliac plexus block
Neurolytic techniques are available to target various specific pain syndromes; the best studied are neurolytic techniques aimed at the coeliac plexus which have been developed to treat severe abdominal pain, most commonly from pancreatic cancer. Although the evidence base is not strong, it is suggestive of benefit, and shows that the procedure is associated with less side effects than management with opioids.  The reviewers noted that less invasive techniques are being developed and should be studied further.
What it means in practice
- Where interventional pain management approaches are available, they need to be incorporated into a holistic plan of care for each patient, and should generally be offered only when optimal medical management is not effective.
- The literature suggests that for carefully selected patients there are potential benefits to be achieved from intraspinal analgesia, and from neurolytic procedures such as coeliac plexus blocks. Sympathectomy has not been shown to be effective.
- Intraspinal analgesia usually requires close ongoing followup from a specialist service to manage implanted devices and medications, and is associated with significant costs and burdens. However when available it may provide an alternative for patients with complex pain that is not well controlled by medical management.
- Vertebroplasty and kyphoplasty are controversial, especially with regard to safety. If offered to patients with advanced malignancy they should not, at this stage, be regarded as minimally invasive.