INTRODUCTION — A substantial number of patients with cancer pain do not obtain satisfactory relief with conventional first-line approaches, including treatment of underlying causes, if possible, opioid-based pharmacotherapy, and noninvasive second-line therapies. For some of these patients, the so called “interventional” pain management strategies may offer safe and effective pain relief. The term “interventional” usually is applied to a group of invasive analgesic therapies, including injection-based treatments, catheter-based infusion therapies, implanted devices, and some surgical approaches.
In the present era, most interventional strategies are nondestructive and are performed using needles (table 1). Some are neurolytic, however, and some involve sophisticated technology, such as implanted neurostimulation and neuraxial drug infusion devices. The evidence base for all of these approaches is limited, particularly in the population with cancer pain, and there are very few controlled trials. Nevertheless, experience in the management of acute and chronic pain suggests that a carefully selected subset of patients with cancer pain may benefit from these procedures.
With the exception of some simple injections (eg, trigger point injections), interventional therapies for pain management are implemented by professionals who have received specialized training. All clinicians involved in the management of cancer pain should have an appreciation for the indications, risks, and benefits associated with the various interventional approaches. If pharmacotherapy is unsuccessful or the benefits versus risks of an intervention outweigh those of other options including pharmacotherapy, an interventional procedure may be a viable alternative, and the patient should have access to a specialist who can confirm the appropriateness of interventional treatments and safely implement them as needed.
This topic will discuss interventional therapies for cancer pain. General principles of cancer pain assessment and management, the use of pharmacologic therapies for cancer pain, and psychological, rehabilitative, and integrative therapies are discussed separately. (See "Assessment of cancer pain" and "Overview of cancer pain syndromes" and "Cancer pain management with opioids: Optimizing analgesia" and "Cancer pain management: Role of adjuvant analgesics (coanalgesics)" and "Cancer pain management: Use of acetaminophen and nonsteroidal anti-inflammatory drugs" and "Rehabilitative and integrative therapies for pain in patients with cancer".)
INJECTION THERAPIES
Soft tissue and joint injections — Injections into soft tissue and joints are often performed in the management of common chronic non-cancer pain syndromes, such as myofascial pain or painful arthropathy. A patient with cancer who develops one of these painful conditions should be considered a candidate for the appropriate intervention, as long as they do not have a contraindication (eg, coagulopathy or leukopenia, pneumothorax), and the decision to proceed or not should be based on a careful assessment of benefits, risks, alternative treatments, and the overall goals of care. (See "Overview of cancer pain syndromes".)
●Trigger point injection – Trigger point injections are used to address focal musculoskeletal pain, including pain that occurs in patients with cancer. Trigger points are distinguished from tender points or regions because they are associated with palpation-induced radiation of pain and taut bands in muscle. In practice, however, injections are often considered for any focal area in muscle or connective tissue that is associated with significant tenderness or pain on motion. These injections are within the scope of routine medical practice and can be performed by nonspecialists.
Although the evidence to support benefit of trigger point injections in cancer patients is limited to anecdotal experience [1-3], this is a reasonable option for patients with cancer pain who have a focal area of myofascial pain and no contraindication to injection, such as a coagulopathy or severe leukopenia. If the pain is primarily myofascial and not secondary to an active, untreated cause, such as metastatic mass or local nerve injury, relief following a trigger point injection often lasts for days or weeks. Dry needling may be performed, or an injection of approximately 1 mL of dilute local anesthetic (such as 1 percent lidocaine or 0.25 percent bupivacaine) may be administered directly into the painful site in the muscle. (See "Complex regional pain syndrome in adults: Treatment, prognosis, and prevention", section on 'Trigger point/tender point injections'.)
●Joint injections – Patients with cancer commonly have painful comorbidities, such as osteoarthritis. Just as injection of a glucocorticoid into a painful joint is widely applied in populations without cancer, it may be considered to treat a painful joint in the cancer population. Again, the decision to proceed should be based on a careful evaluation of risks, benefits, and the goals of care. (See "Intraarticular and soft tissue injections: What agent(s) to inject and how frequently?" and "Joint aspiration or injection in adults: Technique and indications".)
Spine-related injections for back or neck pain — Low back or neck pain may represent important comorbidities in patients with cancer, either related or unrelated to the malignancy. Injections that are commonly performed to address acute and chronic, non-cancer, low back and neck pain may be appropriate in some cancer patients, including epidural steroid injections, facet joint injections, facet denervation approaches, and sacroiliac injections. It is likely that improved survival in some cancers will increase the prevalence of painful spinal disorders in cancer survivors. As a result, the decision to employ the interventional strategies used in chronic non-cancer pain will be encountered with increasing frequency. (See "Management of non-radicular neck pain in adults" and "Acute lumbosacral radiculopathy: Treatment and prognosis" and "Subacute and chronic low back pain: Nonsurgical interventional treatment".)
Vertebral compression fracture — Bone pain is common in metastatic disease, and treatment usually aims to diminish pain, reduce the risk of adverse events, such as fracture, and maintain function. Radiation therapy (RT) and drugs that inhibit osteoclasts may reduce pain and adverse skeletal events; opioids and nonopioid analgesics are administered to address pain. (See "Radiation therapy for the management of painful bone metastases", section on 'External beam radiation therapy' and "Osteoclast inhibitors for patients with bone metastases from breast, prostate, and other solid tumors", section on 'Efficacy and dosing considerations for individual agents' and "Cancer pain management: Role of adjuvant analgesics (coanalgesics)", section on 'Patients with bone pain' and "Cancer pain management with opioids: Optimizing analgesia".)
Vertebral compression fractures in the setting of bone metastases can be very painful and lead to spinal instability. There are several interventional techniques that may be beneficial and potentially opioid sparing.
Vertebral augmentation procedures — Vertebroplasty and kyphoplasty are accepted options for carefully selected patients with symptomatic pathological vertebral fractures without epidural disease or retropulsion of bone fragments into the spinal cord and with pain that is refractory to noninvasive therapies. Vertebroplasty and kyphoplasty are percutaneous injection techniques that may reduce pain and, in some cases, stabilize the fracture. Vertebroplasty involves the percutaneous injection of bone cement (methylmethacrylate) under fluoroscopic guidance into a collapsed vertebral body. Kyphoplasty involves the introduction of inflatable bone tamps into the vertebral body; once inflated, the bone tamps variably restore the height of the vertebral body while creating a cavity that can then be filled with viscous bone cement.
Vertebral augmentation techniques have been used for treatment of painful vertebral collapse in patients with osteoporosis and bone metastases. The evidence in support of vertebroplasty and kyphoplasty for the treatment of painful vertebral collapse in populations with osteoporotic disease is modest and conflicting. (See "Osteoporotic thoracolumbar vertebral compression fractures: Clinical manifestations and treatment", section on 'Vertebral augmentation procedures (vertebroplasty and kyphoplasty)'.)
The evidence is also modest and conflicting in patients with malignant vertebral compression fracture. Nonetheless, systematic reviews of studies conducted in cancer populations report improved pain control in more than one-half of patients, and reductions in analgesic use and pain-related disability scores [4-6]. Although most of the data come from retrospective reports, there is one randomized (although nonblinded) trial of kyphoplasty for treatment in which 134 patients with cancer (37 percent with myeloma) and painful vertebral body compression fractures were randomly assigned to kyphoplasty versus nonsurgical management [7]. Crossover to kyphoplasty was allowed for patients undergoing initial nonsurgical management at one month. In an intention-to-treat analysis, kyphoplasty resulted in decreased back-specific disability at one month and a lower percentage of patients requiring walking aids (46 versus 25 percent), bracing (22 versus 2 percent), bed rest (46 versus 23 percent), and medications of any kind (82 versus 52 percent). All patients who underwent kyphoplasty (whether initially or after crossover from the control group) had sustained improvements over 12 months. There were two adverse events in the kyphoplasty group: one non-Q wave infarction attributed to anesthesia, and adjacent vertebral fracture attributed to the procedure.
Kyphoplasty and vertebroplasty are typically reserved for patients with symptomatic vertebral body fractures without epidural disease or retropulsion of bone fragments into the spinal cord. A vertebral body fracture with a posterior cortical breach is a relative contraindication to kyphoplasty/vertebroplasty. However, at least some data suggest that balloon kyphoplasty is safe and effective in this setting in experienced hands, albeit with a higher risk of cement extravasation [8]. Patients who have involvement of the posterior elements (facet joints or laminae) require an additional posterior tension band. This can be accomplished with the placement of percutaneous pedicle screws at adjacent levels, in addition to cement augmentation at the index level. In the cervical spine, open procedures are still required, although transoral vertebroplasty of the C2 level has been attempted. Gross instability of the cervical and upper thoracic spine requires an open instrumented fusion.
Although clinical experience is limited, American Society for Radiation Oncology (ASTRO) guidelines recommend that RT be used in conjunction with kyphoplasty or vertebroplasty for most patients who have been treated for vertebral metastases [9]. Postprocedure RT for patients with multiple myeloma is controversial and is discussed separately. (See "Multiple myeloma: Overview of management", section on 'Skeletal lesions and bone health'.)
●Contraindications – Vertebroplasty/kyphoplasty should be restricted to patients without epidural disease. Other contraindications include the presence of neurologic damage related to fracture, fractures with a burst component (where bone fragments extend into the spinal canal), systemic or local infection, an uncorrected hypercoagulable state, and severe cardiopulmonary disease [10,11].
●Kyphoplasty versus vertebroplasty – Data comparing kyphoplasty with vertebroplasty in cancer patients are limited. In one small study of 34 patients with symptomatic vertebral compression fractures related to multiple myeloma, patients with >50 percent compression underwent kyphoplasty, whereas patients with <50 percent compression underwent vertebroplasty [12]. Both procedures reduced overall pain and analgesic use, with modestly greater reductions in pain at six months and one year with kyphoplasty. There was no difference in complications.
Given the limitations in the data, the choice between kyphoplasty and vertebroplasty generally is based on clinician preference and expertise, and patient factors. Kyphoplasty is more expensive, but cement extravasation is more common with vertebroplasty. Some practitioners favor kyphoplasty in cases of significant kyphosis (deformity more than 20 degrees) or if there is posterior vertebral cortex involvement, which makes cement extravasation from vertebroplasty more likely [10]. On the other hand, vertebroplasty may be preferred when insertion of the balloon device is technically difficult due to severe vertebral collapse (>65 percent reduction in vertebral height) or if the fracture is more than three months old, in which case elevation of the endplate is unlikely [10].
This approach differs somewhat from the recommendation of an international myeloma working group, which suggests the use of kyphoplasty rather than vertebroplasty for patients with painful vertebral compression fractures [13]. However, a pooled analysis of 34 published case series conducted after this guideline was established concluded that both procedures were equally effective in patients with a vertebral compression fracture related to myeloma [14]. This subject is discussed in more detail elsewhere. (See "Management of complete and impending pathologic fractures in patients with metastatic bone disease, multiple myeloma, and lymphoma", section on 'Kyphoplasty versus vertebroplasty'.)
●Complications – The risk of adverse outcomes appears to be low when the procedure is performed by an experienced physician. Nevertheless, serious complications have occurred, including pulmonary embolism, spinal cord compression, and paraplegia. Intradural cement leakage requiring spinal surgery is a rare complication [5,15].
Alternative therapies — Vertebral fractures may compromise facet structures. For cases in which vertebroplasty/kyphoplasty may not be possible or effective, pain related to the posterior elements may be reduced through medial branch radiofrequency ablation or facet injections [16]. (See 'Spine-related injections for back or neck pain' above.)
NEURAL BLOCKADE — Nerve blocks involve injection of a drug in proximity to a nerve to provide analgesia or anesthesia. Nerve blocks for chronic pain may be diagnostic, prognostic, or therapeutic.
Diagnostic nerve block — A diagnostic nerve block is performed in an effort to better understand the “generator” of the pain or the neural basis for afferent transmission of noxious stimuli. Local anesthetic (LA) blocks of somatic nerves, which interrupt both sensory and motor neural activity, can be useful in localizing the afferent pathway involved in sustaining the pain. As examples:
●A block may distinguish truncal pain that is arising from the abdominal wall from pain that is referred from a visceral source.
●A spinal anesthetic with a sensory level above the level of the patient’s pain can be used to determine whether the pain is centrally or peripherally generated. (See "Spinal anesthesia: Technique".)
If pain persists despite complete interruption of afferent neural input from the painful area, it is plausible to consider a central nervous system (CNS) etiology. In some cases, pain referable to the CNS clearly involves neurologic disease (known as central pain or, more generically, as “deafferentation pain”); in others, it may have a less certain focal cause and may have components of psychological amplification.
In addition to helping to localize the afferent pain pathway, diagnostic blocks may help to clarify whether pain is sympathetically maintained (ie, sustained by efferent activity in a sympathetic nerve or transmitted via afferent nerves [nociceptors] traveling with the sympathetic nerves). Sympathetically-maintained pain is more likely when the clinical findings are consistent with complex regional pain syndrome (CRPS; also called reflex sympathetic dystrophy [RSD] or causalgia). (See 'Sympathetic blocks' below.)
Prognostic nerve block — Prognostic blocks using LA are performed prior to planned neurolytic procedures to determine whether denervation is likely to relieve the pain and whether the sensory loss is tolerable. (See 'Neurolytic blocks' below.)
If a prognostic block with LA does not provide meaningful analgesia for at least the duration of the LA effect, a subsequent neurolysis should not be undertaken. If relief of pain occurs, however, this observation does not guarantee that a destructive procedure will be successful. It is important to explain to the patient that the positive result of the prognostic block supports the subsequent use of neurolysis but does not guarantee that the neurolytic block will work. (See 'Neurolytic blocks' below.)
Therapeutic block — Therapeutic nerve blocks are those that are intended to produce sustained pain relief or to address the underlying anatomic pathway involved in the patient’s pain. These may be broadly divided into those that block nerve transmission without permanently injuring the affected nerve (nonneurolytic) and those that permanently alter the nerve (neurolytic). In general, there is little high-quality evidence comparing these approaches with other analgesic strategies, with the exception of celiac plexus neurolysis (CPN) [17]. (See 'Neurolytic blocks' below.)
Nonneurolytic blocks — Nonneurolytic analgesic blocks can be performed using bolus injection or continuous infusion of an LA. (See "Overview of peripheral nerve blocks" and "Endoscopic ultrasound-guided celiac plexus interventions for pain related to pancreatic disease".)
●Bolus injection – A trial of an LA bolus injection may technically be the same as the diagnostic or prognostic block described previously, but the intent is to produce a meaningful therapeutic outcome, ie, pain relief that far outlasts the pharmacologic duration of drug effect. If this occurs, repeat blocks may be performed to provide sustained pain relief. The reason for prolonged pain relief after temporary disruption of afferent input from a painful site is unknown.
●Continuous infusion – Nonneurolytic blocks may also be performed using an infusion of an LA, which can be delivered via an epidural or perineural catheter. The infusion of LA can be short-term (hours or days) or, occasionally, long-term. Experienced pain specialists may consider perineural LA infusion in selected patients who have a focal regional pain syndrome that is expected to last a short time (but longer than the duration of a bolus injection) and who have demonstrated short-term relief from trials of bolus injection. For prolonged infusion of LA, a catheter can be tunneled under the skin and connected to an external pump. A similar system can be used for neuraxial infusion. (See 'Neuraxial infusion' below.)
While there have been no controlled studies of these techniques for cancer pain, efficacy is supported by many case reports and clinical experience [18-20]. Diverse types of pain syndromes that have not been responsive to more conservative therapies may be amenable. Examples include continuous brachial plexus block via an interscalene catheter for persistent shoulder pain, or continuous epidural infusion for pelvic or lower extremity pain. The open-ended use of a perineural or neuraxial (epidural or intrathecal) LA infusion may be helpful for patients with challenging pain syndromes who have advanced illness and are believed to be near the end of life. In this situation, the potential benefits associated with pain relief may exceed the risk of infection and toxicity from the LA, and the burden and cost of maintaining the infusion may be reasonable trade-offs for pain relief [21]. (See 'Neuraxial infusion' below.)
Sympathetic blocks — Visceral cancer pain reflects afferent input traveling with autonomic nerves, ie, with sympathetic or parasympathetic nerves. Sympathetic nerve blocks interrupt these afferent fibers. The concurrent blockade of sympathetic efferents also may contribute to the positive effects of these blocks, but the mechanisms by which this might occur are uncertain.
When pain originates from a somatic structure, sympathetic block is considered only when there is reason to believe that the pain may be maintained, at least in part, through mechanisms that involve sympathetic efferent activity [22]. Based upon clinical observation, it is widely accepted that the entity referred to as CRPS (also known as RSD or causalgia) is more likely than other painful conditions to have a component of sympathetically-maintained pain. Cancer patients who have regional pain that is associated with focal autonomic dysfunction (such as vasomotor instability or sweating) or trophic changes (such as thinning of the skin, increased hair growth, or brittle, ridged nails) may have CRPS. These characteristics suggest that a trial of a sympathetic nerve block should be entertained. (See "Complex regional pain syndrome in adults: Pathogenesis, clinical manifestations, and diagnosis", section on 'Clinical stages'.)
The efferent sympathetic nervous system originates in the intermediate-lateral horn cells of spinal cord levels T1-L2. The organization of the sympathetic nervous system facilitates isolated sympathetic neural blockade at specific anatomic locations. Common sites for sympathetic blockade include the stellate ganglion, lumbar sympathetic trunk, celiac plexus, superior hypogastric plexus, and ganglion impar. These structures are anatomically separated from somatic nerves, and this distribution allows sparing of sensory and motor fibers during block procedures. It is more difficult to perform segmental blocks of the thoracic sympathetic ganglia without also blocking somatic nerves. Furthermore, the presence of the pleural fold near the sympathetic chain makes a pneumothorax very likely.
Sympathetic block of these structures may be performed using a bolus injection of LA. Following a successful LA injection, a neurolytic block may be performed for cancer patients with advanced disease and intractable pain in the hope of achieving long-term relief. (See 'Neurolytic blocks' below.)
●The stellate ganglion is formed from fusion of the first thoracic and inferior cervical sympathetic ganglion, and it receives innervation from the T1-T4 sympathetic outflow. Stellate ganglion blockade provides sympathetic interruption of the head, neck, upper extremities, and intrathoracic structures. The block is performed by injection of LA near the ganglion where it lies on the anterior surface of the vertebral column, through placement of a needle in the anterior neck.
●The lumbar sympathetic trunk consists of sympathetic fibers that originate in the lower thoracic and L2 segments. Lumbar sympathetic block is performed to relieve pain in the lower extremities. Blockade is classically achieved through a paraspinous approach, with injection at the anterior lateral aspect of the vertebral body on the side of the painful extremity.
●Celiac plexus, superior hypogastric plexus [23], and ganglion impar blocks are performed to relieve visceral abdominal and pelvic pain. They are used to block sympathetic nerves that originate from T5 to L2 and distribute through the greater (T5-T10), lesser (T10-T12), and least (T12-L2) splanchnic nerves, coalescing in the celiac plexus, superior hypogastric plexus, and ganglion impar. These sympathetic nerves also carry visceral afferents, including nociceptors.
Somatic nerve blocks — Bolus injections or continuous epidural or perineural LA infusions can be performed to interrupt afferent input from specific areas of the body [24]. Common types of nonneurolytic somatic nerve blocks include:
●Paravertebral or intercostal blocks interrupt afferent input from a region of the chest or abdominal wall. (See "Thoracic nerve block techniques", section on 'Thoracic paravertebral block' and "Thoracic nerve block techniques", section on 'Intercostal nerve block'.)
●Brachial plexus block interrupts afferent input from the shoulder or arm. (See "Upper extremity nerve blocks: Techniques", section on 'Brachial plexus blocks'.)
●Gasserian ganglion block interrupts afferent input from a part of the face.
●Epidural or intrathecal blocks interrupt afferent input from various areas of the body. (See 'Neuraxial infusion' below.)
Neurolytic blocks — Neurolysis produces analgesia by destroying afferent neural pathways or sympathetic structures involved in pain transmission [25,26]. Both somatic and sympathetic neurolytic blocks may be performed. Neural destruction can be achieved with surgery, cold (cryotherapy) or heat (radiofrequency thermal coagulation), or the injection of a material that damages the nerve (eg, water, hypertonic saline, glycerin, phenol, or alcohol).
All neurolytic techniques result in Wallerian degeneration (ie, degeneration of the nerve axon distal to the destructive lesion) to some degree. However, if the axolemma is intact, nerve regeneration occurs, leading to a return of sensation in approximately three to six months. The extent of the degeneration and the extent and time course of recovery vary with each individual and technique.
All neurolytic interventions carry the risk of producing neuritis or a new “deafferentation pain.” Those methods that preserve neural architecture and allow for regeneration are less likely to be followed by neuritis or a deafferentation pain syndrome. Once a deafferentation pain syndrome develops, it can be as difficult to treat as the original pain for which the procedure was performed. Although uncommon overall, the true incidence of this complication is unknown, and the possibility must be considered when deciding on a therapeutic approach. Patients who undergo neurolysis should be told about this possibility in the informed consent process.
●Indications – Most neurolytic procedures are performed for cancer patients with advanced disease whose pain is not amenable to more conservative therapies [21,27]. In most cases, neurolytic blocks should be considered last resort options because of the risks of late deafferentation pain and unintended neurological compromise, the possibility of unwanted tissue damage induced by the neurolytic process at the time of administration (table 2), and the likelihood that even successful neurolysis may not provide more than a few months of relief. However, there are two major exceptions:
•Celiac plexus neurolysis – CPN is frequently used for pain originating from upper abdominal malignancy, particularly pancreatic cancer. This particular block is sufficiently safe and effective that it is commonly recommended as the next step if one to two trials of systemic opioid therapy are ineffective in relieving pain or if the patient prefers a nonpharmacologic approach to therapy.
The celiac plexus can be accessed intraoperatively, percutaneously using fluoroscopy or computed tomography (CT) guidance, or endoscopically using ultrasound guidance. The decision to employ one or the other typically depends on the discipline and training of the physician. The majority of randomized trials demonstrating benefit from CPN for pancreatic cancer pain in adults have been conducted using the percutaneous approach [28,29]. However, the benefit of endoscopic ultrasound-guided CPN has been shown in at least one randomized trial [30], and this approach is increasingly used as a minimally invasive technique with low risks. (See "Endoscopic ultrasound-guided celiac plexus interventions for pain related to pancreatic disease" and "Endoscopic ultrasound-guided celiac plexus interventions for pain related to pancreatic disease", section on 'Celiac plexus intervention'.)
•Superior hypogastric plexus neurolysis — Superior hypogastric plexus neurolysis may be tried for patients with visceral pelvic pain that is refractory to medical management. The superior hypogastric plexus lies in the retroperitoneum and extends from the anterior aspect of L5 to the superior sacrum. Afferent fibers from the pelvic viscera pass through the plexus, which also contains sympathetic postganglionic fibers. As with celiac plexus block, only visceral pain responds to superior hypogastric plexus block; somatic pain from sacral or muscle involvement and neuropathic pain from nerve root compression or infiltration do not improve. An initial LA block is often used to predict response to neurolytic block [31]. (See 'Nonneurolytic blocks' above.)
The superior hypogastric plexus is accessed intraoperatively or percutaneously via a bilateral posterior approach. Fluoroscopy or CT is used to facilitate needle placement and to confirm the appropriate spread of neurolytic agent.
No randomized studies comparing opioid therapy with neurolysis of the celiac plexus or the superior hypogastric plexus have been published, but several prospective case series indicate that these interventional approaches produce good to excellent pain relief in more than 70 percent of patients with advanced abdominal and pelvic cancers, reduce opioid consumption, and have a favorable adverse event profile [29,31,32]. Although no significant complications occurred in these case series, plausible and significant complications can arise from blockade of the celiac plexus and/or superior hypogastric plexus. Orthostatic hypotension, diarrhea, injury of the aorta, spinal cord, and lumbar plexus, bladder puncture, iliac artery puncture with retroperitoneal bleeding, and cholesterol plaque embolization are some of many potential complications.
Neurolysis also may be considered for selected cancer patients with other types of intractable chronic pain:
•Intractable facial pain – For intractable facial pain, neurolysis (chemical or radiofrequency neurectomy) of the trigeminal nerve branches could lead to deafferentation pain (painful posttraumatic trigeminal neuropathy, also called anesthesia dolorosa) and is generally avoided, if possible. Glycerol injection of the Gasserian ganglion has the lowest risk for adverse effects, the longest duration of activity, and is the preferred approach [33]. For patients who have undergone radical neck/oral surgery, branches of the trigeminal nerve are sacrificed, and the facial pain might be of a deafferentation nature, not nociceptive from tumor. (See "Overview of craniofacial pain", section on 'Painful post-traumatic trigeminal neuropathy'.)
•Medial branch block – Neurolytic blocks of the medial branch of the primary dorsal ramus, which innervate the facet joints of the spine, may be used for treatment of intractable pain and may be beneficial for patients with malignant vertebral compression fractures.
Dorsal punctate midline myelotomy — Dorsal punctate midline myelotomy (DPMM; also called limited transverse dorsal column myelotomy) is a neurosurgical procedure that involves destruction of the midline posterior dorsal column pathway that transmits visceral pain [34,35]. The procedure was developed after the relatively recent recognition of the existence of this pain pathway [36,37]. DPMM may be performed to relieve intractable visceral pain related to abdominal and pelvic cancer.
DPMM appears to be associated with a lower risk of morbidity and complications than earlier procedures (eg, anterolateral cordotomy and commissural myelotomy) [34]. There are no prospective randomized trials comparing DPMM with other interventional pain procedures for cancer pain, and experience with this technique is limited. Case series including small numbers of patients have reported effective pain relief without neurologic morbidity or mortality [35,38-40].
There is growing interest in placing this procedure in a more expansive treatment algorithm for the appropriate patient with intractable cancer pain. However, the place of DPMM in the range of therapeutic options for patients with visceral pain related to abdominal or pelvic cancer remains undefined. A significant issue is that the DPMM procedure is in evolution, with the use of advanced imaging techniques and minimally invasive percutaneous approaches rather than laminectomy, which was initially used for these procedures.
ADVANCED NEURAXIAL TECHNIQUES — Neuraxial techniques include many of the spinal injections discussed above, most commonly including epidural steroid injections. (See 'Spine-related injections for back or neck pain' above.)
Advanced neuroaxial techniques (ie, neurostimulation and catheter-based neuraxial infusion) may be used to provide analgesia without the side effects of systemic pharmacotherapy [41-45]. Proper patient selection is important for the appropriate use of these interventions. The main disadvantages of these techniques are cost, risk of infection, and mechanical failure.
Spinal cord stimulation — The most common type of implanted neurostimulatory treatment is spinal cord stimulation (SCS; dorsal column stimulation). SCS involves percutaneous or surgical implantation of electrodes in the epidural space, with power supplied by an implanted battery. (See "Spinal cord stimulation: Placement and management".)
SCS is not typically used for treatment of pain due to spinal cord disease and is only rarely used for patients with advanced cancer. Although there are no randomized trials addressing the benefit of SCS for cancer-related pain and the quality of the existing data is low [46], it may be an option for patients with neuropathic pain related to cancer or cancer treatment, but not to spinal cord injury, who have failed conservative management [47]. The best candidate for SCS is a patient with intractable focal pain of neuropathic origin (eg, a painful lumbosacral plexopathy or polyneuropathy, or phantom limb pain). The cost of SCS and the need for active patient involvement in its use make it unsuitable for debilitated cancer patients near the end of life.
Neuraxial infusion — Continuous infusion of medications into a neuraxial space (ie, epidural or intrathecal (figure 1)) is also called targeted drug delivery. Several drugs and infusion devices are available for neuraxial infusion. Targeted drug delivery is an important option for patients who have cancer or cancer treatment-related pain and who are refractory to, or intolerant of, systemic pharmacotherapy. Referral to a specialist to evaluate the feasibility and appropriateness of these treatments and others should be considered in the setting of refractory cancer pain.
The decision to try one or another type of neuraxial infusion should be based upon the medical status of the patient, the goals of care, the availability of professional and family support, and cost. The potential for relief of pain when severe and refractory to systemic opioids should be considered when analyzing the risk versus benefit of these procedures. If the priority is comfort, which cannot be achieved with conventional opioid-based therapy, the risk of an intervention may well be worth taking, even in the context of advanced illness or the presence of a relative contraindication (eg, coagulopathy, local infection).
Neuraxial drugs can be infused into the intrathecal or epidural space (figure 1).
Intrathecal versus epidural catheter placement — Epidural delivery permits analgesia to be restricted to fewer dermatomes, which may be an advantage in some patients. However, epidural drug doses are up to 10-fold higher than intrathecal doses, which may increase the risk of systemic side effects (table 3) [48]. There are no randomized trials comparing the efficacy of epidural versus intrathecal drug infusions in patients with cancer pain.
Intrathecal drug delivery is more commonly used for cancer-related pain. Implantable intrathecal infusion systems are exclusively developed to provide targeted drug delivery to the central nervous system (CNS). Medications placed in the intrathecal space are typically concentrated and compounded into specialized regimens. The intrathecal catheter is typically placed below L1, and depending on the location of interest, the radio-opaque catheter tip is advanced under direct visualization via fluoroscopy to a specific spinal level. By way of example, the abdominal and pelvic visceral structures receive innervation from the greater, lesser, and least splanchnic nerves emanating from thoracic levels T5-T12. Exquisite attention to detail regarding placement of the intrathecal catheter tip in the midthoracic levels T6-8 and T9-12 can provide the most effective analgesia for those patients suffering with abdominal and pelvic pain, respectively. It is best to place the catheter at the most optimal anatomical location to cover the patient’s pain.
It is common to place the catheter higher in the intrathecal space than is necessary as the catheter can always be accessed and pulled down, rather than replaced, avoiding repeat and unnecessary dural puncture. In cases where tumor burden obscures the typical lumbar access of the intrathecal space or in the setting of head and neck cancer, epidural rather than intrathecal placement of the catheter may be considered to offer medication delivery to the neuraxial space at the optimal level needed for analgesia.
Both the decision to employ intrathecal versus epidural drug administration and the technical aspects involved in catheter placement are influenced by the patient’s life expectancy:
●For patients with a life expectancy of weeks to months, the catheter may be tunneled under the skin and connected to a subcutaneous port, which is accessed percutaneously and connected to an external pump. In this setting, intrathecal administration may be preferred because of the durability of the catheter placement and the need for lower doses. If an intrathecal catheter is placed, it is often safer and more convenient for the patient to have the tunneled catheter connected to either an external pump or a fully-implanted infusion pump, which is refilled percutaneously. Fully-implanted catheters and infusion pumps are generally considered cost-effective if the patient’s estimated life expectancy is more than three to six months. Both fixed-rate and programable pumps are available, with and without patient control for as-needed bolus dosing. Opioids, local anesthetics (LAs), and other adjunctive medications may be administered and should be added in an algorithmic fashion [49]. The limited capacity of implanted pump reservoirs makes them unsuitable for the high volume of drug solution required for epidural infusion.
A comparison of complications from and some characteristics of epidural or intrathecal long-term administration of opioids for cancer pain is summarized in the table (table 3) [50].
●For patients with a life expectancy of days to weeks, a catheter may be placed percutaneously and connected directly to an external pump. Epidural catheters may be the simpler approach in this setting.
If intrathecal infusion via an implanted pump is under consideration, a temporary trial is often performed first, using single-shot, repeated bolus, or infusion techniques [51]. Management of the patient’s systemic opioid therapy should be coordinated with the pain specialist performing the procedure; systemic opioids may be reduced or eliminated prior to, during, or after the trial is performed.
Choice of agent — A number of medications are used for neuraxial infusion for pain. Opioids (ie, morphine, hydromorphone, and fentanyl) are the most commonly used agents, and they are most often used as monotherapy [52]. Opioids may be combined with LAs (eg, bupivacaine), ziconotide, clonidine, or baclofen to achieve a synergistic effect and reduce the required dose of each medication. There are minimal data comparing different pharmacologic agents, and current regimens have been empirically derived. Information about agents considered useful for neuraxial analgesia for chronic cancer and non-cancer pain, specific recommendations for agents to be used for intrathecal analgesia, and starting doses can be found in the 2017 Polyanalgesic Consensus Conference (PACC) report, which makes recommendations based upon literature review and the consensus of the conference members [49,53]. The most recent evidence-based recommendations for choice of drugs for intrathecal therapy are outlined in the table (table 4) [53].
There are instances in which the medication administered may dictate the location of the intrathecal catheter tip. Many practitioners who utilize the medication ziconotide, an N-type calcium channel blocker with exclusive US Food and Drug Administration (FDA) indication for intrathecal use, will consider high lumbar catheter placement. Ziconotide has the potential to cause psychosis and other CNS-related side effects. To achieve a more local distribution of drug with less rostral spread of the medication, lower catheter placement is thought to be best, regardless of the location of the pain.
Outcomes — There are limited data directly comparing neuraxial analgesia with continued conventional medical treatment in patients with refractory cancer pain [43,54]. The advantages of an intrathecal drug delivery system were shown in a trial that randomly assigned 202 patients with unrelieved cancer pain to intrathecal analgesia or conventional medical management (all pain therapy except spinally administered drugs or other neurosurgical interventions) [43]. Patients assigned to intrathecal therapy were required to have a successful screening trial with intrathecal morphine before pump implantation, and clinical benefit was assessed after four weeks.
At four weeks, both groups had an improvement in pain scores and toxicity, but significantly more patients treated with the intrathecal infusion had a ≥20 percent reduction in the pain visual analog score ((form 1) 85 versus 71 percent), and they had less treatment-related toxicity (58 versus 38 percent). In particular, fatigue and depressed level of consciousness were significantly less common in the intrathecal therapy group. In addition, there was a trend towards longer six-month survival in the intrathecal therapy group (54 versus 37 percent, p = 0.06).
Benefits appear to be sustained. In a later publication with six-month follow-up of this trial, the authors again concluded that the use of intrathecal therapy improved clinical success of analgesic therapy, reduced treatment-related toxicity, and was associated with increased survival [55].
Complications — As with systemic therapy, drugs administered neuraxially may produce side effects and require dose adjustment or change of medication. Respiratory depression is the most serious and should be very rare with proper dosing. Long-term intrathecal infusion of morphine has been associated with catheter tip granuloma. Bupivacaine may cause urinary retention, paresthesias, lower extremity weakness with gait impairment, and occasionally, orthostatic hypotension [56]. Mechanical problems related to the catheter or pump also occur, and the patient must continue to have access to a physician who can manage these eventualities. More serious adverse effects are very uncommon [56-59]. Local infection may occur at the catheter or pump implant site, and meningitis or sepsis is possible but rare. Few intraspinal systems become infected with aseptic technique during placement and with proper maintenance. Bleeding complications and spinal cord injury are similarly rare complications [60].
Specific, comprehensive recommendations for reduction of morbidity and mortality with intrathecal drug infusion have been made by the PACC [61]. The same group has published specific recommendations for prevention, early detection, and treatment of intraspinal granulomas [62].
Intraventricular opioid delivery — Morphine may be injected into the cerebral ventricles after placement of an Omaya reservoir. This technique may be appropriate for certain patients with pain from head and neck cancer, but the available evidence to support superiority of intraventricular over systemic opioid delivery is weak. The relatively small experience with this procedure suggests that 50 to 90 percent of patients obtain good to excellent initial relief [63-65].
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Palliative care" and "Society guideline links: Neuropathic pain" and "Society guideline links: Cancer pain".)
INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, “The Basics” and “Beyond the Basics.” The Basics patient education pieces are written in plain language, at the 5th to 6th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10th to 12th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.
Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on “patient info” and the keyword(s) of interest.)
●Basics topics (see "Patient education: Managing pain when you have cancer (The Basics)")
SUMMARY AND RECOMMENDATIONS
●Interventional therapies may be valuable options for treatment of cancer pain that is refractory to systemic opioids, nonopioid analgesics, and other nonpharmacologic pain control treatments. Any of the generally accepted analgesic procedures performed for patients with acute or chronic pain may be appropriate in some cases of pain related to cancer. With the exception of trigger point injections, all interventional pain procedures reviewed here should be performed by clinicians with specialized training in their use.
●A trigger point injection of a dilute local anesthetic can be used to address focal musculoskeletal pain as long as there is no clinically significant coagulopathy or leukopenia, or relevant risk of pneumothorax. (See 'Soft tissue and joint injections' above.)
●Vertebroplasty and kyphoplasty are accepted options for carefully selected patients with symptomatic pathological vertebral fractures without epidural disease or retropulsion of bone fragments into the spinal cord and with pain that is refractory to noninvasive therapies. (See 'Vertebral augmentation procedures' above.)
A neurolytic block of the medial branch of the primary dorsal ramus may be beneficial for these patients as well. (See 'Neurolytic blocks' above.)
●A diagnostic nerve block may help to localize the afferent pain pathway. (See 'Diagnostic nerve block' above.)
●Nonneurolytic nerve blocks may provide benefit for a diverse array of pain syndromes that have not been responsive to more conservative therapies. Nonneurolytic blocks can be done for either somatic pain syndromes or visceral pain syndromes. (See 'Nonneurolytic blocks' above.)
●Neurolytic techniques produce analgesia by destroying afferent neural pathways involved in pain transmission. With the exceptions of celiac plexus and superior hypogastric plexus neurolysis, these techniques are generally considered “last resort” options. (See 'Neurolytic blocks' above.)
Prior to a planned neurolytic block, a prognostic nerve block using a local anesthetic should be performed; however, relief of pain is not a guarantee of benefit for the subsequent neurolytic procedure. (See 'Prognostic nerve block' above.)
●Dorsal punctate midline myelotomy (DPMM) is a neurosurgical procedure that involves destruction of the midline posterior dorsal column that transmits visceral pain. This is an evolving technique, and experience with its use is limited. (See 'Dorsal punctate midline myelotomy' above.)
●Spinal cord stimulation (SCS) and neuraxial infusion procedures may be used to provide analgesia for selected cancer patients, without the side effects of systemic pharmacotherapy. The main disadvantages are cost, risk of infection, and mechanical failure. Proper patient selection is key to the appropriate use of these interventions. (See 'Advanced neuraxial techniques' above.)
•The best candidate for SCS is a patient with focal pain of neuropathic origin (eg, a painful lumbosacral plexopathy). (See 'Spinal cord stimulation' above.)
•Neuraxial drug infusion should be considered in any patient who has moderate to severe pain that is refractory to conventional opioid pharmacotherapy or in any patient who is intolerant of opioid therapy.
•For patients with a life expectancy of days to weeks, a percutaneously-placed catheter with an external pump is a reasonable option. Epidural placement is usually preferred if life expectancy is short. If survival is likely to be measured in months, the catheter can be tunneled under the skin to reduce the risk of dislodgement and serious infection. In such cases, it is safer and more convenient for the patient to have the tunneled catheter connected to a fully-implanted infusion pump. (See 'Intrathecal versus epidural catheter placement' above.)
•A number of medications are used for neuraxial infusion for pain. Opioids (ie, morphine, hydromorphone, and fentanyl) are the most commonly used agents, and they are most often used as monotherapy. Opioids may be combined with a local anesthetic such as bupivacaine, ziconotide, or clonidine to achieve a synergistic effect and reduce the required dose of each medication. (See 'Choice of agent' above.)
As with systemic therapy, drugs administered neuraxially may produce side effects and require dose adjustment or change of medication. (See 'Complications' above.)
•Intraventricular infusion of morphine after placement of an Ommaya reservoir may be appropriate for certain patients with pain from head and neck cancer, but the available evidence to support superiority of intraventricular over systemic opioid delivery is weak. (See 'Intraventricular opioid delivery' above.)
ACKNOWLEDGMENT — The editorial staff at UpToDate would like to acknowledge Ronald Kaplan, MD, who contributed to an earlier version of this topic review.
