Pain is a complex symptom experienced by many cancer patients. It affects most aspects of life, and controlling it well can make a great difference in patients’ perception of their diagnosis. The etiology of pain in cancer patients is very heterogeneous. Finding the cause directs the treatment and improves the chances of good pain control. In general, cancer pain syndromes can be divided into acute and chronic. Acute ones are usually direct consequence of invasive diagnostic or therapeutic procedures, but they can less commonly be related to cancer itself. Chronic ones are more likely to be caused by the neoplastic process or by antineoplastic therapy.

Metastatic disease may invade bone, obstruct a hollow viscus, and compress nerve or spinal cord. Radiation treatment may cause fibrosis of nerve or spinal cord. Chemotherapeutic agents may cause peripheral neuropathy or aseptic bone necrosis and predispose to painful opportunistic infections. Surgical treatment leads to acute postoperative pain and may cause deafferentation pain if major nerves or nerve plexi are cut. In any given patient, one or more of these factors may be in play, and more than 50% of cancer patients with pain have more than one source of pain.¹

Primary care physicians and oncologists should be able to recognize and treat most cancer-related pain. They should be able to initiate treatment for the more common causes with opioids and nonopioid analgesics. More than 70% of patients can be treated effectively with simple analgesics and adjuvant drugs. Effective pain relief, without intolerable side effects, is occasionally difficult to obtain with the use of conventional analgesics. When this occurs, consultation with a specialist in pain management may be necessary.

Based on rates from 2008 to 2010, National Cancer Institute estimates that 40.76% of men and women born today will be diagnosed with cancer of all sites at some time during their lifetimes.² About half of cancer patients experience pain, most commonly caused by their primary cancer. Pain severity is at least moderate for most patients experiencing cancer-related pain. Pain may also persist in long-term cancer survivors. Cancer-related pain adds to mood disturbance and disability in cancer patients.³

In 1982, Daut and Cleeland found that 36% of 286 patients with nonmetastatic cancer reported pain versus 59% of 381 with metastatic disease.⁴ In 1994, Cleeland and colleagues found that 67% of 1308 outpatients with metastatic cancer had pain, and 62% of those had severe pain. Thirty-six percent reported disability from pain, and 42% of those with pain reported inadequate analgesia.⁵ Terminal pain, refractory to escalating opioid administration, is a more challenging problem. Depression, uncontrolled pain, the adverse effects of opioids, and fear of pain may precipitate suicidal thoughts or requests for aid in dying.^6,7 Pain also adds to the discomfort experienced by those caring for dying patients.

The likelihood of pain associated with cancer depends on the type and stage of disease. Foley, in a 1-week survey of 540 patients hospitalized at Memorial Sloan-Kettering Cancer Center, showed that the prevalence of pain requiring analgesic drugs varies by cancer type (Table 54-1).⁸ In contrast, among 1308 outpatients with metastatic cancer, Cleeland and colleagues did not find variation in pain prevalence according to cancer type.⁵

Gutsgel and colleagues conducted a prospective study on 141 palliative medicine patients. Forty-one patient did not have pain within 1 month before enrollment or could not provide reliable history. Of 100 remaining evaluable patients (71%) who reported pain, 95 had cancer. Sixty-eight percent of them had pain directly caused by tumor, 18% caused by cancer treatment, 3% caused by cancer-associated conditions (e.g., postherpetic neuralgia), and the remaining 16% not related to cancer. Pain was somatic in 52%, visceral in 22%, neuropathic in 10%, and of mixed origin in 16%. Usual pain intensity was moderate to severe in 73%, and maximal pain intensity was moderate or greater in 93% of affected patients. This study supported prior observations of most patients experiencing pain at one site (61%) or two sites (27%) and much less frequently at three or more.⁹

Although significant improvements have been seen in understanding pain and the availability of treatment options, there is still a high prevalence of inadequately treated cancer pain. Both patient and practitioner factors contribute to poor cancer pain assessment and management.¹⁰ Two of three doctors feel insufficiently prepared to manage cancer pain.¹¹ About three of four oncologists and palliative care doctors likewise believe they inadequately assess pain and pain response in their patients.¹² Furthermore, half of doctors are concerned about legal issues related to prescribing opioids for patients with cancer pain, and three of four doctors believe opioid use for cancer pain is associated with high rates of addiction and abuse.^3,11 Patients are often reluctant to report pain. They can have poor treatment adherence, cognitive issues, and affective distress, which can limit reporting, fear of addiction, or developing tolerance and fear of side effects. They often try to be “good patients” by tolerating pain or believe that the doctor should focus on cancer cure rather than pain relief. Some have concerns about negative views of family, friends, and coworkers if they use pain medications.^3,10 There are also considerable gender and ethnic differences in cancer pain reporting and perception, which providers should have in mind when assessing each patient.¹³ All ethnic groups believe that cancer pain is taken more seriously when reported by male patients.^3,13 Meta-analysis of cultural differences in Western and Asian patient-perceived barriers to managing cancer pain has shown significantly higher perceived barriers among Asian patients.^3,14

Pain can be divided into three pathophysiologic categories—somatic nociceptive, visceral nociceptive, and neuropathic—on the basis of the inferred mechanisms of pain (see Chapter 2). It is useful to characterize pain in this way because the approach to treating each type is somewhat different. However, they are not mutually exclusive, and cancer patients, in particular, may have pain with multiple causes. The mechanisms of each type are the subject of considerable ongoing research.

SOMATIC NOCICEPTIVE PAIN

This is the typical pain that we all have experienced acutely or chronically: the cutaneous burn and arthritic joint are examples. The painful site is tender and corresponds to the site of tissue damage. Somatic pain is described as constant and sometimes throbbing or aching. Bone metastases are the most common malignant cause of somatic pain and, in fact, are the most common source of pain in cancer.¹⁵

Noxious (potentially tissue-damaging) mechanical, thermal, and chemical stimuli trigger nociceptive ischemia, inflammation, and perhaps substances produced by a nearby tumor may sensitize nociceptors to ordinarily non-noxious stimuli. Pain signals are carried by small, myelinated Aδ fibers (mechanical and thermal stimuli) and unmyelinated C fibers (all three stimulus types) to the dorsal horn of the spinal cord. From there, they ascend in the contralateral spinothalamic and spinoreticular tracts to the thalamus and reticular formation, respectively (see Chapter 2). Although most research into nociceptive mechanisms has focused on cutaneous pain, nociceptors exist in most tissues to varying degrees.

VISCERAL PAIN

Pain originating from the viscera is familiar to many of us; abdominal cramps and the pain of passing a renal stone are examples. It may be less constant than somatic pain, occurring in dull, colicky waves. Visceral pain is poorly localized and often referred to a distant cutaneous site, which may be tender. Unlike somatic pain, it is often associated with nausea and diaphoresis. Cancer patients experience primary and referred visceral pain from pancreatic cancer, bowel obstruction, and other causes.

The noxious stimuli required to trigger visceral pain include ischemia, inflammation, torsion, traction, distension, and impaction. In fact, cutting, crushing, and burning may not be felt as painful.¹⁶

It took many years to understand that true visceral pain did exist and that somatic tenderness observed during visceral dysfunction was due to convergence of visceral and somatic sensory afferents in the spinal cord, a phenomenon commonly known as “referred pain.” The majority of thoracic and abdominal visceral organs, except the pancreas, are dually innervated by parasympathetic (craniosacral) and sympathetic (thoracolumbar) outflows. Thoracic viscera and upper abdominal viscera are primarily innervated by the vagus (cranial nerve X) and spinal thoracolumbar outflows. The lower abdominal viscera, including the small and large intestine and the urogenital organs, are innervated by thoracolumbar (i.e., lumbar splanchnic nerve and hypogastric nerve) and sacral (i.e., pelvic nerve) outflows.

Sensory afferents innervating the visceral organs are not just a homogeneous group of afferents signaling visceral pain to the CNS. It is now a general notion that pain is primarily signaled by spinal afferents, and vagal afferents signal nonpainful sensations such as hunger, satiety, fullness, and nausea.

Considering the fact that afferent nerve sensitization initiates visceral hypersensitivity, attempts have been made to pharmacologically modulate the excitability of the afferents to alleviate visceral sensitivity. The advantage of targeting visceral afferents with a peripherally restricted drug is to avoid unnecessary CNS complications. Among many target receptors, κ-opioid receptors (KOR), P2X purine receptors, 5-HT₃ and 5-HT₄ serotonin receptors, N-methyl-D-aspartate (NMDA) receptors (NMDAr), tachykinin (NK1, NK2, and NK3) receptors, TRPV1, and γ-aminobutyric acid B (GABA-B) receptors have been documented to have modulating effects on responses of sensory afferents and spinal processing of pain.¹⁷

NEUROPATHIC PAIN

In everyday life, chronic neuropathic pain is uncommon. However, acute and transient neuropathic pain is felt whenever the “funny bone” is struck or an extremity “falls asleep” from pressure. Perhaps the most frequent nonmalignant chronic neuropathic pain is that produced by nerve root compression from a herniated intervertebral disk. Neuropathic pain is often described as prolonged, severe, burning, lancinating, and squeezing and is often associated with focal neurologic deficits. It is usually constant but may be interrupted by paroxysms of dramatically increased pain. There may be no area of tenderness or areas of exquisite sensitivity to normally innocuous stimuli (allodynia). Symptoms and signs of autonomic instability may accompany neuropathic pain. The clinical hallmarks of neuropathic pain are spontaneous pains and painful responses to non-noxious stimuli. Neuropathic pain is also characterized by its relative resistance to opioids, making it the most challenging of pain conditions to treat.¹⁸ Cancer patients experience neuropathic pain from a variety of causes. Direct infiltration of neural structures by tumor is the most common, but iatrogenic causes, such as radiation fibrosis and surgical injury, also occur.

Injury to any part of the nervous system, whether central or peripheral, may result in neuropathic pain. A number of theories have been advanced to explain both peripheral and central mechanisms for maintaining the perceived pain. In the periphery, C fibers may become sensitized by direct injury or by ongoing nociception from injured tissues. This sensitization would make them susceptible to stimuli that are not normally perceived as painful, including sympathetic discharges.^19,20 Two other peripheral mechanisms include neuroma formation (in which an injured nerve’s attempt to regrow results in an overly sensitive, disordered jumble of fibers) and abnormal foci of sensitivity along the course of a nerve, resulting in ephaptic transmission (cross-talk) or ectopic discharges.^21,22

Regardless of the initial site of neurologic injury, the central nervous system probably plays a significant role in maintaining the pain syndrome. Increased sensitivity to peripheral stimuli can be demonstrated in the dorsal horn of the spinal cord, as can spontaneous activity. This spinal cord hypersensitivity could explain both the spontaneous pain and the allodynia felt in neuropathic pain syndromes.

Most pain syndromes in patients with advanced cancer are caused by direct tumor invasion of pain-sensitive structures. The specific pain syndrome depends more on the type of structure involved (e.g., bone pain, visceral pain, mucosal irritation) than on the causative tumor. Similarly, though less commonly, the injury is iatrogenic from diagnostic, surgical, chemotherapeutic, and radiotherapeutic interventions. In many patients, pain has multiple causes. Because increasing pain may signal advancing disease, determining its cause is important.²³ In addition, knowing the cause of pain assists in selecting the most appropriate analgesic approach. The following are several common, recognizable painful conditions that occur in cancer patients. They are grouped according to the type or location of pain.

BONE PAIN

Tumor involvement of bone is the most common cause of cancer pain.¹⁵ Any tumor may involve bone, but the most common include metastatic cancer of the breast, lung, prostate, and thyroid and multiple myeloma.^1,24 This is purely somatic pain unless pathologic fracture or tumor extension disrupts nerve. As such, pain is usually described as focal and constant but may be referred. Typically, patients experience several days or weeks of increasing pain. Acutely increased bone pain may signal fracture or neural impingement. Tumors may activate nociceptors by pressure, ischemia, or secretion of algesic substances (e.g., prostaglandin E₂, osteoclast activating factor).²⁵ Most pain is probably sensed in periosteum and synovium; these are quite sensitive to surgical manipulation. Common sites of bony metastasis are the vertebral column, skull, humerus, ribs, pelvis, and femur.⁸

Hematologic malignancies can rarely produce painful bone marrow expansion, presumably caused by nests of rapidly growing cells in the marrow.^24,26

Diagnosis of bony metastases in known cancer patients may be made by plain radiographs when tumor involves the cortex. A computed tomography (CT) scan further defines the morphology of bone lesions that are seen on radiographs. When the radiographs are normal, radionuclide scintigraphy (bone scan) may identify osteoid formation in the marrow before cortical destruction has occurred. Even in predominantly osteolytic tumors, some reactive osteoid formation usually occurs, and the bone scan result is positive (Table 54-2).²⁷ A bone scan, however, is often normal in purely lytic tumors, such as multiple myeloma, and in previously irradiated bone. Furthermore, it lacks the anatomic detail of radiography. Magnetic resonance imaging (MRI) is more sensitive than radiography and bone scan and can identify bony metastases in previously irradiated bone.⁸ MRI is not used as the initial diagnostic tool because it is expensive, time consuming, and often not immediately available.

Other forms of bone pain in cancer are iatrogenic; these include avascular necrosis of the femoral and humeral heads from steroid treatment, osteoradionecrosis after radiation treatment, and pseudorheumatism from steroid withdrawal. Radiographs do not confirm avascular necrosis for several weeks or months after the onset of pain; a bone scan is more sensitive.⁸ Osteoradionecrosis usually occurs in the mandible and may develop months or years after irradiation. It must always be distinguished from recurrent tumor, radiation-induced sarcoma, and osteomyelitis.²⁸ Reinstitution of steroid treatment, followed by slow withdrawal, confirms the diagnosis of pseudorheumatism by relieving the arthralgias and myalgias.⁸

To relieve the pain, pathologic fractures of long bones are surgically stabilized when feasible. Vertebral collapse is more often treated conservatively with analgesics or with vertebroplasty or kyphoplasty. Radiation adds to pain control of all pathologic fractures. Local field external-beam radiation therapy is an effective palliative modality for painful bone metastases, with pain relief seen in 80% to 90% of cases.²⁴ A rare paraneoplastic form of renal phosphate wasting called oncogenic osteomalacia can cause osteomalacia, multifocal bone pain, and fractures.^24,29 This rare syndrome is most often associated with mesenchymal neoplasms, and complete tumor removal can lead to rapid correction of the biochemical derangements, remineralization of bone, and symptom improvement.

BACK PAIN

The vertebral column (particularly the thoracic spine) is the most common site of bony metastases.³⁰ Although cancer causes less than 1% of back pain in the general population, 98% of known cancer patients who present with back pain have underlying malignancy.³¹ Up to one-third of cancer patients develop metastases to the spine, with prostate, breast, thyroid, and lung cancers being most common.^30,32 Because back pain in cancer patients usually signifies bone or epidural metastasis, aggressive investigation to define the presence and extent of tumor is necessary. Left untreated, metastases destabilize the axial skeleton and encroach on the spinal cord or cauda equina.

Investigation of back pain should begin with a detailed history and physical examination, attending to the presence of rapid pain progression, referral patterns, and neurologic symptoms and findings. Points that should raise the suspicion of epidural disease include failure of pain to resolve with recumbency; point tenderness of the spine on examination; and, of course, any history of bowel or bladder dysfunction or focal neurologic deficit. Vertebral disease at certain levels of the spine may initially have a confusing presentation. High cervical spine metastases may produce only posterior headache, which could be mistaken for tension headache. Involvement of C7 to T1 causes pain in the interscapular region. Lesions of T12 or L1 may refer pain to the flank, iliac crest, or sacroiliac joint. Sacral destruction may refer pain in a saddle pattern.¹ Radiation myelopathy causes local burning pain, which radiates bilaterally, and progressive neurologic deficits.³³

Stable back pain in cancer patients, without neurologic symptoms or signs, warrants nonurgent radiography of the affected area. Radiographs will detect approximately 70% of vertebral tumors.³¹ If plain radiographs are normal, radionuclide bone scan is indicated because it has a higher sensitivity than radiography for early osteoblastic lesions, fractures, and infection. If radiography or bone scan is positive or equivocal, MRI should be performed to detect or define the extent of disease, especially that involving soft tissue.²⁷ If both radiography and bone scan results are normal, CT or MRI of the paraspinal and retroperitoneal areas is warranted to detect a source of referred or extraaxial soft tissue pain.

Magnetic resonance imaging is always the test of first choice when there is evidence of neural compression and should be completed urgently if cord compression is suspected. MRI is sensitive for early metastases, easily images the entire spine in one scan (unlike CT), and accurately defines the extent of adjacent soft tissue disease.²⁷ CT with myelography may be used when MRI is not available.

Even when there is no evidence of neural compression by history or physical examination, radiologic evaluation should not be delayed for more than several days if back pain is positional or progressing. Unstable vertebral fractures may lead to acute cord compression and may require prophylactic stabilization or irradiation.³⁴

Some patients with back pain due to epidural malignancy do not have vertebral metastases. Tumor may reach the epidural space by hematogenous spread or by direct extension along nerves, through the intervertebral foramina. Neoplastic epidural spinal cord compression (ESCC) is a common complication of cancer that can cause pain and potentially irreversible loss of neurologic function. Most often, ESCC is caused by posterior extension of a vertebral body metastasis into the epidural space. ESCC almost always presents initially as back or neck pain. Because pain usually precedes neurologic impairment by weeks or months, it is crucial to diagnose epidural disease extension in patients when pain is the sole complaint so that effective treatment may be started to prevent or retard the progression of neurologic impairment. For most patients, MRI is the preferred approach to evaluate the epidural space.²⁴

For most patients, radiation therapy (RT) represents first-line definitive treatment for ESCC. Glucocorticoid treatment is useful to temporarily improve pain and neurologic functioning, often providing a window of time during which RT can be provided. Surgical decompression is considered if the tumor type is typically radioresistant and the lesion is high grade, the neurologic status is deteriorating during RT, ESCC occurs in a previously irradiated field, the lesion is posterior and can be easily extirpated, or a tissue diagnosis is needed.²⁴

Other patients with back pain in the absence of vertebral disease may have leptomeningeal carcinomatosis (LC). They have signs of neurologic dysfunction at several levels and usually complain of headache, nausea, and nuchal rigidity but may also have lumbar radicular pain.³³ Contrast-enhanced MRI of the entire spine and head followed by lumbar puncture is appropriate when LC is detected. LC may require treatment with radiation, steroids, or intrathecal chemotherapy.

Loss of motor function, hyper- or hyporeflexia, and bowel or bladder disturbances are suggestive of myelopathy. Their presence should prompt immediate intervention, even before diagnostic imaging has been obtained, to prevent permanent neurologic impairment (Fig. 54-1). Even in the absence of myelopathy, certain situations require urgent radiologic evaluation. Rapidly progressing pain is highly suspect for tumor and should be investigated until the extent of disease is known. Increased pain when supine or erect may signal positional cord compression, which will progress to overt myelopathy unless treated. In these situations, MRI of the entire spine should be performed to define the extent of disease and the threat to neural tissue.³⁵