NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Jacox A, Carr DB, Payne R. Management of Cancer Pain. Rockville (MD): Agency for Health Care Policy and Research (AHCPR); 1994 Mar. (AHCPR Clinical Practice Guidelines, No. 9.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of Management of Cancer Pain

Management of Cancer Pain.

Show details

5 Nonpharmacologic Interventions: Invasive Therapies


40. With rare exception, noninvasive treatments should precede invasive palliative approaches. (Panel Consensus)

41. Indications for palliative radiation therapy include treatment of symptomatic metastases in sites where tumor infiltration has caused pain, obstruction, bleeding, or compression. (B)

42. Radiopharmaceuticals emitting a b-particle should be used for the pain of bone metastases only when bone scintigraphy shows a lesion. (A)

43. Radiation tolerance of adjacent normal tissues should be considered in the design of treatment portals and the prescription of teletherapy or radiopharmaceutical dose. (A)

44. The desired dosage of radiation should be administered in the fewest fractions possible to promote patient comfort during and after treatment. (Panel Consensus)

45. Neurolytic blockade of peripheral nerves should be reserved with rare exception for instances in which other therapies (palliative radiation, TENS, pharmacotherapy) are ineffective, poorly tolerated, or clinically inappropriate. (Panel Consensus)

46. Clinicians should:

  • Assess thoroughly each patient's pain mechanism in order to apply the most appropriate nerve block.
  • Screen patients for coexistent medical conditions, ability to understand risks of the proposed procedure, and ability to cooperate during the procedure.
  • Consider a block only if the person planning to do it is experienced and skillful, is prepared to deal with its immediate effects and side effects, and is able to provide follow-up assessment and treatment.
  • Use radiographic control for blocks when ease and safety depend on precise anatomic guidance. (Panel Consensus)

47. When a patient is painfree after neurolysis, opioids should not be stopped abruptly, lest a withdrawal syndrome be provoked. (B)

48. The oncologic surgeon should be familiar with the interactions of chemotherapy, radiation therapy, and surgical interventions so that iatrogenic complications may be avoided or anticipated. (Panel Consensus)

49. The surgeon should recognize and treat characteristic pain syndromes that follow specific surgical procedures. (Panel Consensus)


Invasive nonpharmacologic interventions complement behavioral, physical, and drug therapies in a substantial minority of patients in whom these therapies alone do not control pain (see Figure 3). With rare exception, noninvasive analgesic approaches should precede invasive palliative approaches. Although radiotherapy and surgery can cure primary disease, they are discussed here in relation to pain relief only. Surgical procedures are useful in selected patients to debulk tumors and hence reduce symptoms of obstruction or compression. Anesthetic and neurosurgical methods can be used to ablate pain pathways or implant devices for drug delivery or electrical stimulation of neural structures. For any invasive therapy, the risks, availability of expertise and suitable support systems, and cost should be considered in addition to the apparent necessity or medical indication.

Radiation Therapy

Radiation therapy can relieve metastatic pain as well as symptoms from local extension of primary disease (Greenwald, Bonica, and Bergner, 1987). Over one-third of the practice of radiation therapy is palliative (Arcangeli, Micheli, Arcangeli, et al., 1989). The intent of any palliative treatment is to relieve pain quickly and maintain symptom control for the duration of the patient's life. Treatment is therefore tailored to the patient's clinical condition and overall prognosis (Lawton and Maher, 1991; Maher, Coia, Duncan, et al., 1992). Radiation therapy is complementary to analgesic drug therapies and may enhance their effectiveness because it directly targets the cause of pain.

In general, the larger the daily dose of radiation, the lower the total dose that can be administered because of limits to normal tissue tolerance. Proportionately more tumor cells are killed when the daily radiation dose is larger. A balance is required between the killing of tumor cells and the adverse radiation effects on normal tissues, which are largely a function of the daily dose. A number of different schedules have been developed that take into account specific tumor characteristics and the tolerance of normal tissues. The literature is divided regarding the optimal radiation schedule to achieve tumor regression (Hall, 1993; Thames, Withers, Peters, et al., 1982) and disease palliation (Price, Hoskin, Easton, et al., 1986) at either primary or metastatic sites. Generally, however, radiation treatment is planned in relation to clinical status.

The toxicity of radiation is determined by the structures included within the radiation portal, the dose per fraction, the total dose, and the radiation sensitivity of the tissues involved. The desired dosage of radiation should be administered in the fewest fractions possible to promote patient comfort during and after treatment. Radiation side effects are restricted to the radiation portal and can be classified as either acute, occurring during or immediately after the course of radiation therapy, or late, occurring months to years later. Acute radiation effects are more prominent with radiation schedules that deliver high total doses of radiation with small daily fractions; they generally begin at the end of the second week of therapy (Hall, 1993). Acute radiation effects, occurring primarily at skin and mucosal surfaces, usually consist of an inflammatory response such as skin erythema or pigmentation, or as mucositis. Acute reactions, however, are generally mild during palliative therapy schedules, which deliver high daily radiation doses over 1 to 2 weeks. Late radiation effects may arise without any preceding acute reactions. Fibrosis is the most common type of late radiation injury and can be observed in many types of tissue, including skin. Late effects are most prominent in tissues with limited regenerative capacity such as brain, peripheral nerves, and lungs. Because of the limited duration of patient survival, however, late effects are seldom seen after palliative radiotherapy.

Bone Metastases

Most patients referred for palliation of metastatic bone pain have primary tumors of the highest overall incidence: breast, prostate, or lung. Other neoplasms involving bone, such as myeloma, also respond to radiation. Regular followup should be provided to offer treatment of new symptoms or the use of other palliative techniques such as radiopharmaceuticals.

Plain radiographs are useful in detecting lytic or blastic lesions from bone metastases. Bone scintigraphy, however, is more sensitive than skeletal radiography for the detection of most bone metastases. Although 73 percent of patients in one series were asymptomatic when skeletal metastases were discovered by scintigraphy, 66 percent of these symptom-free patients ultimately experienced moderate to severe bone pain (Sherry, Greco, Johnson, et al., 1986a). In patients who experience bone pain and have normal bone scans, MRI may be a helpful diagnostic tool.

Indications for the radiation of bone metastases include pain relief and the prevention or promotion of healing of pathologic fractures. Spinal cord compression associated with vertebral collapse due to bony or epidural metastases requires emergent radiation therapy, sometimes in coordination with surgical intervention to preserve neurologic integrity (Bates, 1992). Orthopedic complications, including pathologic fracture and spinal cord compression, have been reported in 36 percent of breast cancer patients with skeletal metastases (Sherry, Greco, Johnson, et al., 1986b). Lytic lesions that are 2.5 cm or larger in weight-bearing bones or that cause a more than 50 percent loss of cortical bone place patients at high risk for pathologic fracture (Bates, 1992); patients with such lesions may benefit from prophylactic surgical fixation in conjunction with adjuvant irradiation.

Pain Relief With Localized Radiation Therapy.Radiation is commonly administered to a localized bone metastasis. An analysis of therapeutic results is complicated by variation in the location and extent of bone metastases, primary histology, individual differences (including patients' underlying medical conditions), and co-administered treatments. Concurrent analgesic use is frequently a confounding, poorly quantified variable in many accounts of pain control during local radiation of a metastasis. Most retrospective and prospective studies report that 75 percent or more of patients obtain relief from pain and that about half of those who achieve relief become pain-free (Nielsen, Munro, and Tannock, 1991). However, selection bias cannot be excluded; valid and reliable pain assessment instruments were not commonly used.

The literature is divided on appropriate fractionation (Blitzer, 1985; Hoskin, 1988; Tong, Gillick, and Hendrickson, 1982). Protracted regimens of more than 10 treatments may be more appropriate for patients with life expectancies of longer than 6 months to reduce potential late radiation effects or acute effects such as nausea if critical structures such as the stomach have to be included in the radiation field. For patients with a more limited life expectancy, radiation can be administered in fewer fractions, depending on the patient's clinical status (Lawton and Maher, 1991; Maher, Coia, Duncan, et al., 1992). These later regimens result in effective palliation in over 70 percent of patients at 3-months' followup, with negligible complications when radiation portals are localized (Arcangeli, Micheli, Arcangeli, et al., 1989; Bates, Yarnold, Blitzer, et al., 1992; Blitzer, 1985; Tong, Gillick, and Hendrickson, 1982).

Wide-Field Radiation Therapy.Hemibody irradiation, which can treat multiple disease sites, is particularly appropriate for diffuse bone pain. A single large fraction of 6 Gy to 8 Gy is administered to one half of the body. If necessary, the other half can be treated after a 3-week interval to allow for bone marrow recovery. With antiemetics and partial shielding to reduce lung exposure, toxicity occurs in fewer than 10 percent of patients, and 50 percent experience stabilization of disease at 1-year followup (Poulter, Cosmatos, Rubin, et al., 1992). Salazar, Rubin, Hendrickson, et al. (1986) reported that palliation was achieved in 73 percent of patients treated with hemibody irradiation, and pain recurrence was lower than that reported in an earlier uncontrolled study of the palliative effects of local radiotherapy (Tong, Gillick, and Hendrickson, 1982). This analysis is consistent with other reported studies of hemibody irradiation in which 50 percent of patients report at least partial pain relief within 48 hours of treatment with an eventual total response rate of 55 to 100 percent (Kuban, Schellhammer, and el-Mahdi, 1991; Salazar, Rubin, Hendrickson, et al., 1986).

Radiopharmaceuticals.Several radiopharmaceuticals have been used therapeutically. Iodine-131, used for the treatment of multiple bone metastases from thyroid cancer, results in bone scan evidence of response in 53 percent of patients (Maxon and Smith, 1990). Phosphorus-32-orthophosphate has provided partial or complete relief of pain in about 80 percent of patients with bone metastases from breast and prostate carcinoma ( Silberstein, Elgazzar, and Kapilivsky, 1992). In an analysis of 18 published studies, strontium-89 was found to provide partial to complete pain relief for 65 percent (Silberstein, unpublished manuscript). For example, Silberstein and Williams, (1985)) reported a palliative response of 51 percent of patients, and Robinson, Spicer, Preston, et al. (1987) reported 80 to 89 percent palliative response. In these studies, analgesic use and activities of daily living were used as measures of palliation. Myelosuppression, manifested by approximately a 30 to 50 percent decline in leukocyte and platelet levels within 4 to 6 weeks, generally occurs in patients with either extensive disease or pretreatment peripheral cytopenia (Lewington, McEwan, Ackery, et al., 1991). Rhenium-186 and samarium-153 phosphonate chelates have demonstrated 65 to 80 percent efficacy in international clinical trials, with FDA approval pending (Maxon, Schroder, Thomas, et al., 1990; Turner, Claringbold, Hetherington, et al., 1989). These beta-emitting radiopharmaceuticals, which require only a single intravenous injection, are used to relieve pain from widespread, osteoblastic skeletal metastases visualized with bone scintigraphy. If pain recurs, 50 percent of patients will respond to a second administration.


Painful nerve compression or infiltration by a malignant tumor can sometimes be alleviated by radiation therapy. These primary tumors often require fractionated radiation therapy over 5 to 7 weeks in an attempt to secure local or regional control of the disease. Dosage is limited by the proximity of the tumor to radiosensitive structures, such as the spinal cord. Peripheral nerves, however, can tolerate higher doses.

Other Therapeutic Applications

Palliative radiation can be administered to any location of symptomatic primary or metastatic disease. Aggressive, sometimes protracted multimodality therapy may be given to patients with certain primary tumors, such as soft tissue sarcomas and carcinomas of the breast, lung, and rectum, to relieve both symptoms and to achieve control of advanced disease. Palliative radiation may also be given to metastatic lesions involving the brain, eye, skin, and soft tissue. Localized radiation may be used to treat lymph node involvement causing symptoms due to pressure on adjacent nerve roots and blood vessels. Intra-abdominal tumors may infiltrate the retroperitoneum and adjacent nerve roots or may cause local symptoms such as bowel obstruction. Although limited by the tolerance of the bowel to radiation, some tumor regression and symptomatic relief may be accomplished through fractionated radiation. Because the radiation tolerance of normal liver or kidney is even lower than that of bowel, treatment of pain due to capsular distention of either organ is rarely undertaken. Radiotherapy is generally not administered in these cases unless a trial of analgesic therapy and, when appropriate, chemotherapy has been unsuccessful. Symptomatic bleeding from endobronchial, cervical, and bladder tumors can often be stopped by external beam irradiation.


Brachytherapy involves the placement of a radioactive source within tissue to deliver localized radiation and is frequently applied to treat recurrent disease in an area previously treated by external beam radiation. Advantages include the sparing of critical structures close to the tumor, and brevity of treatment (hours to days). Difficulties primarily involve anatomic constraints on implant placement. Common applications include the endoluminal treatment of recurrent endobronchial and bile duct tumors, the intracavitary treatment of cervical and endometrial cancer, and interstitial implants in unresectable tumors with catheters or radioactive seeds. Occasionally, hyperthermia will be combined with either brachytherapy or external beam irradiation to relieve pain and other symptoms of recurrent disease originating from head and neck or breast cancers.

Anesthetic Techniques

Nerve Blocks

The possibility of controlling otherwise intractable pain by the relatively brief application of a local anesthetic or neurolytic agent makes neural blockade an attractive approach in selected patients. Published estimates of the percentage of all patients with cancer pain for whom nerve block procedures may appropriately be considered vary greatly. Variability in this estimate reflects evolution of the effectiveness of noninvasive therapies, interinstitutional differences in availability of clinicians with the necessary expertise, and access to alternative options such as spinal opioid therapy or neurosurgery (Bonica, Buckley, Moricca, et al., 1990). Allowing for vagueness in methods of arriving at published estimates, lack of uniformity in clinical conditions treated by neural blockade, and in reported clinical outcomes, it still appears that some 50 to 80 percent of patients who receive nerve blocks for cancer pain may benefit (Cousins and Bridenbaugh, 1987; Patt, 1993; Raj, 1992) (Table 19).

Table 19. Nerve Blocks.


Table 19. Nerve Blocks.

Local anesthetic such as lidocaine or bupivacaine is typically applied at an anatomically defined site to provide diagnostic information (e.g., whether the pain is somatic or visceral; whether it has a sympathetic mechanism). Prognostic injection assesses side effects such as hypotension and subjective sensations, including pain relief or unpleasant numbness, likely to result from a planned neurodestructive procedure. Although the lack of a desirable result from local anesthetic injection after proper needle placement generally predicts the failure of a neurolytic block, a promising result after local anesthetic injection does not guarantee the success of subsequent chemical destruction.

Therapeutic injections of a local anesthetic may provide relief that outlasts its pharmacologic action. Prolonged benefit may follow injection of trigger points for myofascial pain -- a procedure sufficiently simple, safe, and efficacious that it can be accomplished by many primary care providers. The injection of an anti-inflammatory corticoid with a local anesthetic into the spinal space or around nerve roots can reduce edema and irritation produced by tumor compression and provide analgesia for days to weeks. One or more cervical or lumbar sympathetic blocks may result in prolonged relief in patients whose cancer-related pain is sympathetically maintained. Sympathetic block performed during acute herpes zoster infection can immediately decrease pain, hasten resolution, and avert the development of postherpetic neuralgia, and should be considered as preemptive therapy for this debilitating sequel (Ferrer, 1989). The simpler technique of subcutaneous infiltration with local anesthetic and corticoid has also been reported to provide symptomatic relief for herpes zoster. When single sympathetic blocks produce only transient benefit, the placement of a catheter at the sympathetic ganglion (or the corresponding intraspinal segments or interpleural space) to enable continuous sympathetic blockade for days to weeks may produce sustained benefit.

Patient selection and timing of neural destruction for pain relief are based on the exhaustion of more conservative modalities, a lack of available, clinically superior options, and the availability of capable physician and support systems after the procedure. Nondestructive analgesic infusion techniques can preempt the need for neurolytic procedures. Therapeutic choices depend on patient and family preferences and the clinical judgment of their health care providers (Verrill, 1990).

Peripheral nerve destruction can be accomplished by the injection of ethanol, phenol, or other neurolytic agents at sites where a previous test injection of local anesthetic has produced pain relief. Whereas phenol induces warmth and then numbness, alcohol produces intense transient burning after injection and hence should be immediately preceded by local anesthetic injection. Small volumes of alcohol or phenol may be injected intrathecally to destroy nerve root function in a localized distribution. Approximately 60 percent of patients treated with intraspinal alcohol or phenol experience complete or near-complete relief of pain until death (Rodriquez-Bigas, Petrelli, Herrera, et al., 1991). When a patient is painfree after neurolysis, opioids should not be stopped abruptly, lest a withdrawal syndrome be provoked. Complications including paresis, paralysis, and bowel or bladder dysfunction affect 0.5 to 2 percent of patients treated with intraspinal alcohol or phenol ( Gerbershagen, 1981). An epidural injection of phenol (or alcohol, according to some reports) can accomplish the same goal; however, the targeting of the injectate is less precise, the neurolytic effects take place over a more diffuse area than that affected by the intrathecal route, and the technique is less well established than intrathecal injection (Salmon, Finch, Lovegrove, et al., 1992).

Neurolytic sympathetic blockade is useful to relieve pain in the arm, head and neck (stellate ganglion), or leg (lumbar sympathetic block), as well as to interrupt the visceral afferent pain pathways mediating pain in the pancreas and other upper abdominal organs (celiac block) or in the pelvis (hypogastric block). Side effects of celiac block include transient hypotension and diarrhea; complications (less likely with radiologic guidance) include paraplegia or less severe radicular weakness or numbness, intrarenal injection and damage, retroperitoneal hematoma, and failure of ejaculation (Ischia, Ischia, Polati, et al., 1992; van Dongen and Crul, 1991). Four-fifths or more of patients with pancreatic or other abdominal cancers derive pain relief from celiac block, usually lasting until death (Brown, Bulley, and Quiel, 1987; Eisenberg, Carr, and Chalmers, unpublished manuscript; Mercandante, 1993). Even when relief is incomplete, patients may appreciate the ability to lower their opioid dosage and by doing so reduce drowsiness and constipation. It thus appears reasonable to consider early celiac neurolytic block for patients with a short life expectancy and pain from pancreatic cancer (Mercadante, 1993). A recently reported technique for refractory chest wall tumor pain is interpleural blockade, which uses long-term local anesthetic infusion or single-dose phenol (Lema, Myers, de Leon-Casasola, et al., 1992).

Neurolytic blockade of peripheral nerves should be reserved for instances in which other therapies (palliative irradiation, TENS, pharmacotherapy) are ineffective, poorly tolerated, or clinically inappropriate. Suitable targets for this approach include intercostal nerves at the site of painful tumor, after maximal doses of radiation and systemic analgesics, or nerves of the head and neck (e.g., gasserian ganglion). Pain recurrence due to neuritis is common because an alcohol-damaged nerve regenerates over weeks to months. If the mechanism of pain is partial or complete denervation, this will not be corrected (and may potentially be worsened) by further chemical damage to the nerve.

Pain that is diffuse (e.g., from multiple bony metastases) may respond to chemical ablation of the pituitary, which is accomplished by alcohol administered through a needle advanced transnasally until its tip rests in the pituitary fossa (see also Neurosurgery, below). Pain relief by this intervention may be rapid and striking, while ascending nociceptive pathways remain unharmed. Pain relief has been reported in about two-thirds of patients, whether or not the primary tumor is hormone dependent (Takeda, Fujii, Uki, et al., 1983). Complications include headache, persistent leakage of CSF, coma, and cranial nerve palsies, all of which occur at a frequency of 5 percent or less (Cook, Campbell, and Puddy, 1984). Diabetes insipidus is a predictable side effect of complete pituitary ablation.

Technical aspects of the above procedures are beyond the scope of this guideline and are well described in a number of recent monographs (Abram, 1989; Charlton, 1986; Cousins and Bridenbaugh, 1987; Swerdlow, 1987). Complications associated with local anesthetic nerve blocks, catheter implants, neurostimulator implants, thermal ablations, and neurolytic injection have been reported (Cousins and Bridenbaugh, 1987; Melzack and Wall, 1990; Raj, 1992). Serious side effects including hemorrhage, infection, unexpected nerve damage, pneumothorax, and cardiorespiratory arrest are rare but nonetheless mandate resuscitative skills and close short-term followup.

Because of the appeal of nerve blocks for use in intractable pain and their potential for harm as well as benefit, clinicians should:

  • Assess thoroughly each patient's pain mechanism, in order to apply the most appropriate block.
  • Screen patients according to coexistent medical conditions (e.g., coagulopathy); ability to understand risks of the proposed procedure (e. g., paresis or incontinence); and ability to cooperate during the procedure (e.g., not move).
  • Consider a block only if the person planning to do it is experienced and skillful; prepared to deal with its immediate effects and side effects (e.g., hypotension, respiratory depression, or paralysis); and able to provide followup assessment and treatment.
  • Use radiographic control for blocks when ease and safety depend on the precise identification of landmarks.

Catheter Placement for Drug Delivery

Temporary spinal or epidural catheter placement is normally undertaken by specialists trained to recognize possible complications (e.g., opioid-induced respiratory depression or hypotension or sensorimotor blockade due to local anesthetic) and able to deal with these promptly and effectively. The need for dosage titration and coordination of spinal with systemic medications and nonmedical therapies requires that the catheter be placed within the framework of multidisciplinary continuing care. Because identical materials and methods are often used for percutaneous epidural catheter placement for cancer pain and for acute postoperative pain control, anesthesiologists typically perform these techniques and their specific followup. Factors to consider are presented in Table 12. The placement of catheters other than spinal ones, such as for drug infusion into interpleural or paravertebral areas, is uncommon, and few data other than case reports are available.

Percutaneous electrical stimulation for the relief of otherwise refractory cancer pain has likewise not yet been evaluated in controlled trials. Case reports -- limited essentially to the percutaneous insertion of spinal cord electrodes for dorsal column stimulation -- tend to focus on details of the method, to use nonuniform patient selection criteria, and to use heterogeneous pain assessment methods and followup duration. Not all experience is favorable ( Meglio, Cioni, and Rossi, 1989). Hence, as Miles and colleagues wrote nearly 20 years ago, "At this stage it seems sensible to concentrate effort on evaluating the method rather than on encouraging widespread and possibly indiscriminate use of what is an expensive use and relatively unproven technique" (Miles, Lipton, Hayward, et al., 1974).


Neurosurgical procedures for the relief of pain include neuroablation, implantation of drug infusion systems, and neuroaugmentation. Published estimates of the percentage of patients who require neurosurgical procedures to control cancer-related pain vary but are nearly all fewer than 10 percent. Nevertheless, long-standing clinical experience supports a view that neurosurgical intervention is appropriate for patients in whom more conservative treatment is neither tolerated nor effective (or is unlikely to be effective) and for whom the expertise and followup care are available. The choice of procedure is based on the location and type of pain (somatic, visceral, deafferentation), the general condition of the patient, the life expectancy, and the expertise available.

There is no simple, entirely safe procedure to alleviate cancer-related pain. Depending on the clinical setting and procedure, the risks of neurosurgical operation include new pain symptoms from nerve damage at the site of incision or nerve division, recurrence of pain after a transiently successful result, and postoperative neurologic impairment. These risks must be balanced against an ideal possible outcome of abolition of pain with little or no need for medication. In a particular clinical situation, a lack of personnel with experience in carrying out and following up other invasive therapies may warrant greater reliance on neurosurgical options (e.g., cordotomy instead of epidural catheter for pain of pelvic tumor). Because appropriate patient selection is essential, each proposed neurosurgical intervention is best reviewed by a team of oncologists, pain specialists, psychotherapists, and neurosurgeons.

The following discussion addresses only procedures that are in general use and for which reported results can be meaningfully assessed. Classic (White and Sweet, 1969) and recent textbooks and monographs provide current reviews of all of these procedures (Bonica, 1990; Gybels and Sweet, 1989; Patt, 1993).


Peripheral Neurectomy.Currently, peripheral neurectomy for the control of cancer pain has largely been supplanted by other techniques, such as neuraxial opioid infusion or lytic nerve block. Multilevel neurectomy for chest wall pain is indicated when a discrete pain-producing lesion can be demonstrated to involve several intercostal nerves (Arbit, Galicich, Burt, et al., 1989). Neurectomy may also be effective in alleviating pain originating from a paraspinal tumor that involves a nerve or nerves at or distal to the neural foramen; it is often performed at the time of an operation on the spine (e.g., anterior or posterolateral vertebrectomy). Cranial neurectomies have selected indications in neuralgias resulting from cancer. The trigeminal and glossopharyngeal nerves can be ablated by radiofrequency lesions created by electrodes placed in either the foramen ovale (gasserian ganglion) or the jugular foramen or by chemical neurolysis at the gasserian ganglion (Giorgi and Broggi, 1984; Ischia, Luzzani, and Polati, 1990; Sweet, 1976).

Dorsal Rhizotomy.Selective ablation of the dorsal nerve root reduces nociceptive perception in the affected area and spares motor function. Multilevel dorsal rhizotomy of all roots supplying an extremity leads to a functionless limb. The likelihood of this impairment is lessened by sparing one dorsal root. In practice, this procedure is considered only for localized pain in the trunk or abdomen or, rarely, for an extremity that is functionless preoperatively (Arbit, Galicich, Burt, et al., 1989; Sindou, Fischer, Goutelle, et al., 1981). Dorsal rhizotomy can be accomplished by chemical neurolysis with radiographic guidance to place the tip of an infusion catheter at the precise segment within the epidural space. Surgical rhizotomy may be necessary if expertise in chemical neurolysis is unavailable or if it has been tried unsuccessfully.

Anterolateral Cordotomy (Spinal Tractotomy).Anterolateral cordotomy is an ablative procedure aimed at the pain-conducting tracts in the anterolateral quadrant of the spinal cord. Cordotomy provides selective loss of pain and temperature perception several segments below and contralateral to the segment at which the lesion is placed. Anterolateral cordotomy is effective for unilateral, mainly somatic pain below the midcervical dermatomes (Ischia, Ischia, Luzzani, et al., 1985; Lahuerta, Lipton, and Wells, 1985). For visceral pain or bilateral pain, bilateral cordotomies may be required (Amano, Kawamura, Tanikawa, et al., 1991). Most cordotomies are currently done with the patient under local anesthesia by the percutaneous route under fluoroscopic guidance, and the lesion is created by radiofrequency. The percutaneous approach avoids risks of open operation and anesthesia in patients in poor medical condition.

Open cordotomies require a laminectomy and are most frequently performed at the low cervical or upper thoracic spine. Open cordotomy may benefit patients in whom a percutaneous procedure has failed, those who cannot cooperate because of severe pain or confusion, those at risk for respiratory compromise, or those with bilateral pain in whom a bilateral, high cervical cordotomy carries additional risk of neurologic impairment. Potential complications include unmasking of dysesthetic pain; bladder, bowel, and sexual dysfunction; ataxia; paresis; and sleep apnea (Lahuerta, Lipton, and Wells, 1985; Tasker, 1988).

Commissural Myelotomy.Commissural myelotomy disrupts pain-conducting fibers as well as a polysynaptic pain pathway that runs through the center of the spinal cord. Indications for myelotomy are bilateral and midline pelvic or perineal pain (Adams, Lippert, and Hosobuchi, 1988; van Roost and Gybels, 1989). The procedure may produce sphincter or motor dysfunction. Open myelotomy involves a multilevel laminectomy and exposure of the appropriate lumbar or sacral segments of the spinal cord. By use of an operating microscope, a midline incision is made and the spinal cord is divided vertically (Gildenberg, 1984). A cervicomedullary junction (extralemniscal) myelotomy that is performed stereotactically with CT guidance, local anesthesia, and intraoperative physiologic assessment can achieve pain relief over wide areas of the body including midline structures (Schvarcz, 1978). Potential complications include temporary dysesthesia and limb apraxia (Gildenberg, 1984).

Hypophysectomy.Surgical and chemical (stereotactic transsphenoidal) hypophysectomy are similar procedures that each offer a 40 to 70 percent likelihood of pain relief (Levin, Katz, Benson, et al. , 1980). The mechanism by which pain relief is achieved is unknown, but it is not related directly to the expected fall in the pituitary hormone levels, because pain relief is achieved in hormonally independent and dependent tumors (Katz and Levin, 1977; Lipton, Miles, Williams, et al., 1978; Takeda, Fujii, Uki, et al., 1983). The clearest clinical indication is for bilateral or diffuse bone pain from metastatic disease that has failed to respond to all other hormonal, radiation, or medical therapies. Hormone replacement therapy is needed to replace pituitary secretion. Potential complications include endocrine deficits, damage to the optic nerves or oculomotor apparatus from the injected chemical agent, and CSF leakage (Cook, Campbell, and Puddy, 1984; Lahuerta, Lipton, Miles, et al., 1985; Lipton, Miles, Williams, et al., 1978).

Neuraxial Opioid Infusion

In properly selected patients, intraspinal or intraventricular infusions of opioids have the advantage of producing profound analgesia without motor, sensory, or sympathetic blockade ( Behar, Magora, Olshwang, et al., 1979; Bullingham, McQuay, and Moore, 1982). See Chapter 3 and also the section on catheter placement in this chapter for a discussion of intraspinal and intraventricular routes of administration.


Interest in endogenous pain control systems as a therapeutic target began over 20 years ago in the context of the contemporaneous discoveries of the positive reinforcing quality of electric self-stimulation of the brain in animals and humans. Profound analgesia without drugs was reported in laboratory animals during electrical stimulation of the brain stem (Reynolds, 1969; Yaksh and Rudy, 1976). These effects appear to depend on the body's own opioids, endorphins. Since then, electrical stimulation for cancer pain control has been directed at deep brain structures such as the periaqueductal and periventricular grey areas (Meyerson, Boethius, and Carlsson, 1978; Young and Brechner, 1986), the limbic system (Gol, 1967), and other more superficial sites such as the pituitary gland (Yanagida, Suwa, Trouwborst, et al., 1988). Evaluation of the efficacy of electrical stimulation ("neuroaugmentation") of deep brain structures for cancer pain relief, as for many other modalities, is difficult because of scanty descriptions of patients' diagnoses and limited pain assessment and followup, as well as the relatively few patients treated in this fashion compared with much larger numbers treated, e.g., with pharmacotherapy. Nonetheless, the few descriptive, uncontrolled published studies report partial or complete pain relief in 27 to 76 percent of patients treated by neuroaugmentation (Meglio and Cioni, 1982; Meyerson, 1982; Young and Brechner, 1986). Published results of spinal cord stimulation for cancer pain relief are less encouraging. Meglio and others (Meglio, Cioni, and Rossi, 1989) reported, in a series of 109 patients treated for pain relief by means of spinal cord stimulation, that none of the 11 who had cancer pain derived any clinical benefit, in contrast to favorable responses observed in patients with vasculopathic pain or postherpetic neuralgia. Similarly, others (North, 1993; Gybels, 1993; Marchand, 1993) have found spinal cord stimulation useful to treat chronic pain if not due to malignancy and only anecdotal observations support the success of this modality in patients with cancer-related pain (Miles, Lipton, Hayward, et al., 1974; Nittner, 1980; Raj, 1992).


Operations for the curative excision or palliative debulking of a tumor have the potential to reduce pain, improve prognosis, and even to achieve long-term, symptom-free survival. On the other hand, a tumor may be recognized to be unresectable at the time of operation. These perioperative dilemmas provoke anxiety in patients and their families, who worry not only about mortality but also about possible survival at the expense of function or loss of body parts. This anxiety may worsen pain.

The surgeon's response to these issues can help to create a sense of personal comfort, to reduce the feelings of loss of control in patients confronted with a loss of autonomy, if not life itself, and to foster a clear understanding of the pain- and tumor-control goals of the surgical procedure and of how the procedure relates to other aspects of treatment.

Postoperatively, the patient is often left with major changes in anatomy and physiology (e.g., laryngectomy, colostomy) that require further rehabilitation and continued attention to pain control. The surgeon should convey the nature and implications of the surgical intervention to the other members of the patient's management team and should continue in an advocacy role throughout the patient's course of care (Dunphy, 1976).

Continuing surgical care is ideally provided in the context of an interdisciplinary approach with an attempt to avoid fragmentation and duplication. A vital part of surgical care for malignant disease is followup to contribute to improving quality of life, particularly the reduction of pain and suffering, to give assistance in rehabilitation, and to provide psychological support for the patient ( O'Young and McPeek, 1987). The patient and family seek assurance that the surgeon will not abandon them postoperatively and will be available for acute and chronic pain control.

Surgical Management of Pain Due To Primary or Metastatic Tumor

Pain control is usually a secondary goal when curative tumor excision is performed. In contrast, when surgery is palliative because the tumor is unresectable, pain control is frequently the operative indication. Cancer pain may arise through a variety of mechanisms that are amenable to relief by surgery. During curative or palliative procedures surgeons should use techniques to limit the development of chronic neuropathic pain such as nerve-sparing incisions, avoidance of ischemia, and careful dissection around nerves. The oncologic surgeon should be familiar with the interactions of chemotherapy, radiation therapy, and surgical interventions so that iatrogenic complications may be avoided or anticipated (e.g., multiple fistulas resulting from bowel resection performed after radiation).

Surgical procedures should proceed within a framework of basic surgical oncology principles that relate to both curative and palliative surgery. Among these is the fundamental principle that the first resection offers the best opportunity for cure. Occasionally, an anatomic structure will need to be removed during this initial curative procedure to achieve a resection margin free of tumor. Careful preoperative and intraoperative weighing of the benefits of potentially curative procedures versus the risks of chronic pain and disability should be done with the patient preoperatively and confirmed by the surgeon intraoperatively.

The second principle is that even outstanding radiotherapy generally will not improve an inadequate surgical procedure. Depending on the natural history of a specific tumor system and anatomic constraints, postoperative radiotherapy can enhance local tumor control in patients with microscopically positive margins. However, radiotherapy may not provide long-term durable control of residual macroscopic gross malignant disease. Therefore, a surgical procedure undertaken for pain control is also most effective and durable if all gross disease can be resected. Short- and long-term pain control may be enhanced with postoperative radiotherapy. As a corollary to this principle, simple tumor debulking generally does not provide durable palliation for most patients. Growth of the residual tumor is sufficient to cause a rapid recurrence of painful symptoms. In such patients, radiotherapy alone may provide better palliation than an incomplete resection with its attendant pain.

The third principle is that local tumor recurrence is not always the harbinger of disseminated disease. For example, local recurrence of soft tissue sarcoma or colorectal carcinoma does not always behave in this manner. Consequently, some patients with local recurrence can still be cured, depending on their underlying disease. Such patients may be best served by a second resection aimed at cure, perhaps incorporating radiotherapy, rather than a less aggressive palliative procedure solely for pain control.

The fourth principle is that the timing of a surgical intervention for pain control is important. Understanding the natural history of a tumor system includes an awareness of the potential of a given tumor type for local invasion as well as tumor-specific patterns of metastasis. This addresses the issue of prospective palliation in which the effectiveness of surgical pain control may be maximized if undertaken before the onset of symptoms. For example, stabilizing a long bone with lytic metastases where fracture is likely or decompressing a spinal canal to prevent impending paralysis or to relieve tumor nerve root entrapment may be most helpful as a timely intervention before the development of irreversible symptoms. This is in contrast to the asymptomatic patient with a recurrent, widely metastatic carcinoma of the colon that may eventually cause intestinal obstruction. In these instances, it may be better to monitor the patient closely for the development of symptoms rather than intervene early and disrupt the quality of life.

Pain as a Consequence of Operation

Surgical procedures can cause several different forms of pain, including incisional pain. Depending on the resection and the specific tissues removed, patients may experience deep wound pain that may be more difficult to control. Finally, many patients may experience a variety of chronic pain syndromes after surgery (see Table 5). Some of these may not emerge until weeks or months after discharge. The surgeon should recognize and treat characteristic pain syndromes that follow specific surgical procedures (e.g., mastectomy, nephrectomy, etc.).

Careful surgical technique frequently can ameliorate the severity of postoperative pain. Gentle tissue handling; use of nerve- or vessel-sparing procedures; avoidance of tissue ischemia; careful neurolysis, performed as needed with a dissecting microscope; and selection of non-muscle-splitting incisions can contribute to less painful surgery and recovery.

In the postoperative period, the surgeon should encourage the full use of the pain-control armamentarium. The management of patients with acute pain such as that caused by pathologic fracture, surgery, or diagnostic treatment procedures is described in detail elsewhere (Acute Pain Management Guideline Panel, 1992).

The surgeon should assess the quality of postoperative pain control by frequent, direct patient contact. The surgeon without expertise in pain management should seek consultative help, particularly for the treatment of special populations. An integrated multidisciplinary pain control approach will maximize the usefulness of surgery as an adjunct of pain control in the patient with cancer.


  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...