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

Goudas L, Carr DB, Bloch R, et al. Management of Cancer Pain. Rockville (MD): Agency for Healthcare Research and Quality (US); 2001 Oct. (Evidence Reports/Technology Assessments, No. 35.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.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


Overview of the Evidence Report

This report summarizes the scientific evidence on several key questions related to cancer pain formulated by a panel of technical experts see (Appendixes A and B). We review the published literature on the epidemiology of cancer pain and its relief and summarize predominantly randomized controlled trials on the efficacy of common treatments. Patients with cancer often experience pain from causes unrelated to cancer, and treatment of such pain cannot be omitted from their care (Abram, 1989; Cousins and Bridenbaugh, 1998; Doyle, Hanke, and MacDonald, 1999; Foley, 1999; Foley and Payne, 1989; Jacox, Carr, Payne, et al.,1994; Modell 1961; Paris, 1997; Patt, 1993; Payne, Patt, and Hill, 1998; Portenoy and Lesage, 1999; Raj, 1992; Swerdlow, 1986; Task Force on Pain Management, 1996). In this literature review we considered cancer-related pain as that caused by the disease itself or by its treatment, such as surgery, radiation therapy, or chemotherapy. Patients with postmastectomy pain were included, as were patients with pain resulting from the side effects of antitumor treatment, such as mucositis. We did not, however, include trials strictly concerned with the treatment of acute postoperative pain nor trials that studied the use of analgesics for control of symptoms other than pain, e.g., dyspnea (Bruera, MacEachern, Ripamonti, et al., 1993).

The topic of cancer-related pain was selected by the Agency for Healthcare Research and Quality (AHRQ, formerly Agency for Health Care Policy and Research (AHCPR)) in response to a request from the American Pain Society. In 1997, AHRQ designated 12 institutions in the United States and Canada to serve as Evidence-based Practice Centers (EPCs). The EPCs prepare evidence reports and technology assessments on topics selected by the AHRQ that focus on specific aspects of prevention, diagnosis, treatment, or management of a particular condition or on an individual procedure, treatment, or technology. These evidence reports and technology assessments are based on rigorous, comprehensive, and systematic reviews of the scientific literature and on explicit, detailed, and documented methods, rationales, and assumptions. They often include meta-analyses and cost and decision analyses. While evidence reports provide systematic reviews of the literature on specific key questions, they are not intended to be compendia of knowledge about the designated areas.

All EPCs collaborate with other medical and research organizations in developing these reports and thereby obtain input from a broad range of experts. Professional associations, health plans, providers, and others that nominate topics may act as partners with EPCs, providing technical expertise and serving as peer reviewers of the final product. EPC evidence reports are not intended as surveys of current practice or as clinical practice guidelines. The present report provides background information and summaries of evidence for use by different groups, including primary care practitioners; nurses; pharmacists; physical therapists; specialists in oncology, pain treatment, or other disciplines; and policy decision makers. The partners -- not the EPCs -- are expected to apply the findings from the evidence reports and technology assessments to support development of clinical practice guidelines or other implementation tools to improve the quality of care in their respective organizations. Thus, these evidence reports and technology assessments provide an evidence-based foundation for public and private organizations to develop their own clinical practice guidelines, performance measures, review criteria, or other clinical quality improvement tools. In addition, these reports may give health plans and payers the information needed to make informed decisions about coverage policies for new and changing medical devices and procedures.

What is cancer Pain?

Careful clinical observations of the experience of pain by patients with cancer indicate that "cancer pain" is a nociceptive mosaic composed of acute pain, chronic pain, tumor-specific pain, and treatment-related (including procedure-related) pain, cemented together by ongoing psychological responses of distress and suffering (Bonica, 1985; Chapman and Gavrin, 1999; Loeser and Melzack, 1999). The metaphor of cancer pain as a mosaic conveys the emergence of a single, unified whole from many separate pieces assembled across time (Loeser, 2001).

Although it is a reminder of cancer-related mortality and carries profound personal, social, cultural, and religious implications (Institute of Medicine, 1997; Cassel and Foley, 1999), pain associated with cancer shares mechanisms with acute or chronic noncancer pain. Traditional definitions of chronic pain require that it be present a minimum of 3 to 6 months (Fields, 1995; Task Force on Pain Management, 1997; Aronoff, 1999; Ashburn and Staats, 1999). Yet current pain research (Besson, 1999) confirms that every physiological feature considered essential for chronic pain -- central sensitization, hyperalgesia, novel gene expression, synaptic remodeling ("plasticity"), "pain memory" formation, and behavioral adjustment -- is triggered within days of acute, ongoing tissue injury (Carr and Goudas, 1999). A new pain symptom that leads to the diagnosis of cancer rarely is eradicated before it has the opportunity to provoke acute physiological responses that may initiate chronic pain. When a new pain appears in a patient known to have cancer, it may still remain noticeable even when diminished by analgesic therapy and hence persist with sufficient intensity to sustain discomfort in an already-sensitized nervous system.

Space-filling noncancerous lesions, such as abscesses, herniated intervertebral discs, or benign adenomas, cause pain through the local release of diverse inflammatory mediators and by exerting pressure on surrounding tissues, including nerves (Ashburn and Rice, 1999; Wall and Melzack, 1984). Inflammatory mediators associated with cancer include prostaglandins, cytokines, tumor necrosis factors (Sorkin, Xiao, Wagner, et al., 1997) interleukins, growth factors, and other tumor-derived products such as endothelin (Davar, 1998), each of which can excite nociceptors (Schwei, Honore, Rogers, et al., 1999). Some cancers induce endogenous antibodies, and others are treated by therapeutic administration of exogenous antibodies; both classes of agents are capable of producing painful neuropathies just as may be seen in autoimmune or infectious diseases (Start, Yu, Yakash, et al., 1997; Sorkin, 2000). Nontumorous causes of viscus obstruction, such as ureteral stones or bowel adhesions, can also cause pain, sleeplessness, nausea, and distress. Destructive lesions from infection, such as osteomyelitis, or enzymatic damage, as in pancreatitis, readily induce pain Staats in (Payne, Patt, and Hill, 1998). Preclinical research in an animal model of bone cancer pain has revealed a distinctive neurochemical and histological "signature" in afferent nerves and their spinal cord connections (Honore and Moutyh, 2000). Yet given the spectrum of possible pain mechanisms (Table 1) it is clear that several elements within it may be active in a single patient with noncancer or cancer pain (Woolf, Bennett, Doherty, et al., 1998). If several pain mechanisms exist simultaneously in many patients with cancer pain, one may ask whether a mechanism-based treatment algorithm is necessarily more effective than a pain-intensity-driven one. In the latter approach, each patient is considered potentially to have both nociceptive and neuropathic pain components and receives therapy titrated according to pain intensity (see below).

Table 1. Categories of pain and related mechanisms (adapted from Woolf, Bennett, Doherty, et al., 1998).


Table 1. Categories of pain and related mechanisms (adapted from Woolf, Bennett, Doherty, et al., 1998).

Both cancer-related and noncancer pains may involve neuropathic components, in which the nervous system is damaged (Woolf and Mannion, 1999), and nociceptive components, in which injury to nonneural tissue is conveyed through an undamaged nervous system. Mindful that some noncancerous causes of chronic pain can degrade quality of life as much as, if not more than, cancer pain (Aronoff, 1999; Committee on Pain Disability and Chronic Illness Behaviour, 1987; Loeser and Egan, 1989) many pain clinicians and patients now urge that the term "nonmalignant pain" be abandoned. Present-day practical therapeutic options for cancer pain relief overlap substantially with those for noncancer pain (see Section 1.5), with the exception of a few strategies applied to the latter but not the former, such as the use of biphosphonates, radionuclides, and chemical or surgical neurolysis (see Figure 1).

Figure 1. Relationship between "generic" and "cancer" pain treatment options.


Figure 1. Relationship between "generic" and "cancer" pain treatment options.
NOTE: The larger circle denotes techniques broadly applied to treat many forms of pain regardless of site, chronicity, or etiology. Such (more...)

Although common features of noncancer and cancer-related pain are plentiful, distinctive aspects of the latter deserve emphasis because of their clinical implications for patient counseling and therapy (Kanner, 1988). First, pain from cancer (as for any condition) tends to increase in severity with advancing disease -- a tendency worth remembering when planning hospice and terminal care (Kane, Bernstein, Wales, et al., 1985; Ganzini, Nelson, Schmidt, et al., 2000). Recognition that an increase in the severity or frequency of pain may herald disease progression or recurrence is useful in clinical decision making (Jacox, Carr, Payne, et al., 1994).

Second, patients with cancer often experience pain at multiple sites concurrently, through multiple mechanisms, and with distinct patterns, such as continuous pain, movement-related pain, and spontaneous breakthrough pain (Ashby, Fleming, Brooksbank, et al., 1992). Addressing only one source and type of pain may be inadequate.

Third, clinicians have identified a number of cancer pain syndromes, some of which are tumor-specific patterns of local or distant metastasis whereas others reflect diffuse neuropathies from tumors or chemotherapy (Payne and Gonzales, 1999). Clinicians caring for patients with cancer are expected to have the ability to recognize (and when feasible, anticipate) these syndromes and to respond to them promptly. For example, increasing back pain and radiculopathy may herald impending spinal cord compression by epidural metastasis, which must be promptly evaluated and treated to avert neurological catastrophe. In contrast, the subacute onset of arm pain after an otherwise successful course of mantle radiation need not prompt a relentless search for a tumor because this symptom is consistent with a radiation-induced brachial plexopathy. The expectation that an invasive procedure will provoke brief but intense pain should elicit preemptive action, such as local anesthetic pretreatment at the site of venipuncture or lumbar puncture. Syndrome recognition and specific medical management, while crucially important for clinical diagnosis and therapy, fall outside the scope of the present evidence review (Cherny, Chang, Frager et al., 1995; Cherny and Portenoy, 1999; Foley, 1985; Portenoy and Lesage, 1999).

Fourth, the nervous system itself may become dysfunctional as a result of cancer, through mechanisms such as brain metastases, meningeal carcinomatosis, or a paraneoplastic syndrome, or through a secondary tumor effect, such as inappropriate antidiuretic hormone secretion causing hyponatremia, sedation, and exaggerated somnolence in response to analgesic medication (Meyers 2000) Tumor-related cognitive impairment can interfere with pain assessment and confound analgesic titration. Last, as emphasized throughout this literature review, the total experience of chronic cancer and noncancer pain encompasses more than just pain intensity; it also includes family, spiritual, and financial dimensions (Ferrell and Ferrell, 1996; Lang and Patt, 1994).

The Epidemiology Cancer and Cancer-related Pain

Cancer has a profound impact on public health throughout the world as a result of its prevalence and devastating morbidity and mortality (Little, 1998;Wolf in Parris, 1997; World Health Organization [WHO], 1993). Pain control addresses only one dimension in the global degradation of quality of life that patients with cancer may suffer, but it is a key dimension (Cleeland, 1991; Ahmedzai, 1995). Intractable pain remains one of the complications most feared by patients with cancer, both in itself and as a harbinger of global loss of control and finally, mortality (Jacox, Carr, Payne, et al., 1994; Greenhalgh and Hurwitz, 1998). In the United States alone in 1999, the American Cancer Society estimates that some 560,000 people died from cancer -- more than 1,500 people per day (Landis, Murray, Bolden, et al., 1999). More than 1.2 million Americans were newly diagnosed with cancer in 1999, according to National Cancer Institute and American Cancer Society estimates.

Because nearly 10 million Americans now have or previously had cancer, it contributes substantially to the national burden of disease, whether measured in terms of mortality, disability-adjusted life-years, or hospital days (Gross, Anderson, and Powe, 1999). Despite substantial and rising funding from the National Institutes of Health for clinical and basic cancer research, age-adjusted incidence rates aggregated across age, race, and cancer type have increased during the past 25 years (Landis, Murray, Bolden, et al., 1999; National Cancer Institute, 1999). During this interval, the incidence of cancer rose 19.9 percent among Whites, 24.3 percent among African Americans, and 20 percent among all races. All of these increases are statistically significant.

As the population of the United States has grown, its age distribution has shifted to include a higher proportion of elderly people (U.S. Bureau of the Census, 1999). The aging of the American population parallels a global trend throughout developed countries, in which the elderly are expected to become a sizable burden on their countries' economies (Mitchell, 1997; Public Health Service, 1990). The impact of cancer increases dramatically with age (American Cancer Society, 1998). Although cancer is the second leading cause of death (after accidents) in children younger than 14 years in the United States, 5-year survival rates have improved substantially for many childhood cancers since the 1970s (Landis, Murray, Bolden, et al., 1999). Alleviating cancer pain in children must be assigned a high priority (Schechter, Berde, and Yaster, 1993), and many children who die of cancer have substantial suffering at the end of life (Wolfe, Grier, Klar, et al., 2000). Current pooled population estimates place the cancer incidence rate according to age of diagnosis at less than 50 cases per 100,000 in those under 25, a figure that rises steadily to over 200 per 100,000 in the 40-44 age group. Pooled data for all ages from birth to 54 yield an incidence of just over 100 cases per 100,000; this rate increases 10-fold to over 1000 cases per 100,000 in the 55-64 age group, and further doubles in those over 65 to greater than 2000 cases per 100,000 (National Cancer Institute, 1999). Correspondingly, the cumulative percentage of the U.S. population experiencing invasive cancers during their lifetime increases sharply from below 2 percent in the 0-39 age group, to just under 10 percent in the 40-59 age group, to about 30 percent in those older than 60 (see Figure 2) (Landis, Murray, Bolden, et al., 1999; National Cancer Institute, 1999).

Figure 2. Actual and projected growth of the United States population and incidence of cancer, overall and age 65 years and older.


Figure 2. Actual and projected growth of the United States population and incidence of cancer, overall and age 65 years and older.
SOURCE: SEER Cancer Statistics Review 1973-1996, National Cancer Institute; U.S. Bureau of (more...)

The combination of an expanding and aging population, an increase in cancer incidence pooled across all diagnoses and ages, and a sharp age-related upsurge in cancer incidence and prevalence in those over 60 guarantees that the national disease burden of cancer will grow further during coming decades and that this burden will continue to fall disproportionately on the elderly (Ferrell and Ferrell, 1996). Work in the 1970s and 1980s by Bonica (1985), Twycross (1976), Foley (1985), Daut and Cleeland (1982), and Stjernsward (WHO, 1993) established that about three-quarters of patients with advanced cancer experience pain. Surveillance data from developed countries around the world include data from about 58,000 patients. This global experience (see Chapter 3 and Evidence Table 1) indicates that one-third to one-half of all patients undergoing active cancer treatment experience pain and that the likelihood of pain is influenced by type of tumor, stage of disease, and extent of metastases (Bonica, 1985; Daut and Cleeland, 1982).

Assessment of Cancer Pain

Evaluation of the patient's total pain experience is the foundation of a comprehensive diagnostic and therapeutic plan (Fields, 1995; Spross, McGuire, and Schmidt, 1990a, 1990b). The subjective nature of pain makes its assessment indirect and often challenging, yet the very process of exploring this aspect of a patient's personal, internal experience affirms the importance of this experience (Morris, 1998) and demonstrates a patient-centered point of view (Gerteis, Edgeman-Levitan, and Daley, et al., 1999). In contrast, much of medical care, including cancer care, has in the past suffered from a disease-centered focus. Pain assessment in cancer care should extend beyond nociceptive evaluation to consider comorbid medical and psychosocial problems, the meaning and impact of pain on the patient and significant others, and its effect upon quality of life (WHO, 1993).

A survey of developmentally and culturally appropriate instruments to assess pain and quality of life is outside the scope of the present literature synthesis, given its focus on pain treatment. However, the current consensus is that initial evaluation of any patient with cancer-related pain must include the essentials expected of any pain history (location, intensity, quality, temporal characteristics, exacerbating and relieving factors, and responses to prior treatments), together with psychosocial assessment, physical examination, and appropriate diagnostic studies (Jacox, Carr, Payne, et al., 1994). The goal of evaluation of the nociceptive dimension of the pain experience is to establish a pathophysiological mechanism for each pain, whenever possible, as a clinical syndrome whose key features, natural history, and optimal therapy are well recognized (Caraceni and Portenoy, 1999; Portenoy and Lesage, 1999). Clinical assessment as to whether pain is neuropathic, nociceptive, somatic, or visceral (Gebhart, 1995) can influence initial selection of drug or nondrug therapy such as surgery or radiation therapy. However, neuropathic pain is not a unitary entity and results from deafferentation, mono- or polyneuropathies, or a complex regional pain syndrome (Woolf and Mannion, 1999). Nociceptive pain may result from somatic or visceral damage (Cervero and Larid, 1999). Recently, persuasive arguments have been advanced that this traditional, anatomical classification may be less useful than a mechanistic classification (Woolf, Bennett, Doherty, et al., 1998). Table 1 illustrates that pain from tissue or nervous system injury may involve a variety of coexisting mechanisms, each of which may be a target for specific pharmacological, surgical, or physical therapy interventions.

Unless cancer pain intensity is assessed systematically using a validated scale, it is difficult to judge the benefits, or lack thereof, of any analgesic regimen, let alone to compare one regimen with another (Jadad, 1994; Max, 1996; McQuay and Moore, 1998). The 0-10 visual or verbal analog scales, or variants thereof such as a thermometer, are validated and easy to administer. Their use is common but by no means universal in clinical trials of cancer pain relief. Patient self-report is more accurate than vital signs, outward behavior, or observer estimates. Because patients experience pain in diverse ways (Erbman, 1934) and are often reluctant to complain of pain, studies that infer the severity and quality of pain solely on the basis of patient complaint or chart review are biased toward underestimating its incidence, prevalence, and severity (Daut and Cleeland, 1982). To reduce such bias and increase the accuracy and precision of pain measurement, clinicians must assess pain prospectively (McCaffery and Pasero, 1999).

Treatment Options for Cancer Pain

This section surveys the major current options for cancer pain treatment, without intending to present these as treatment recommendations. It is intended to provide a background for interpretation of the results of systematic reviews (see Chapter 3) and related conclusions (see Chapter 4). Systemic pharmacotherapy, principally with oral agents, is the foundation for treating cancer pain because of its relative low risk and cost, dependability, and ease of administration (Jacox, Carr, Payne, et al., 1994; American Pain Society, 1999). However, nondrug measures are also important, and cognitive-behavioral interventions offer dual benefits for pain control and coping. Because patients differ in their acceptance of and responses to specific analgesics or adjuvants, and to different behavioral strategies, it is essential that treatment be individualized (McQuay, 1999; Warfield 1993).

The three principal families of drugs used to manage cancer pain are NSAIDs or acetaminophen, opioid analgesics, and adjuvant analgesics. Adjuvant drugs treat concurrent symptoms that exacerbate pain (e.g., insomnia), enhance the analgesic efficacy of opioids, or provide analgesia for specific types of pain (e.g., neuropathic pain) and include antidepressants, anticonvulsants, and antiemetics. Drugs from these three principal families are often given in combination.

A simple, widely applied approach to managing cancer pain, developed by the WHO, is the "three-step analgesic ladder" (or "staircase"). The first tier, for mild to moderate pain, consists of NSAIDs and acetaminophen with or without adjuvant medications. As pain escalates or persists, treatment progresses to the second tier, in which a weak opioid, such as codeine or hydrocodone, is added to the NSAID with or without an adjuvant drug. Discussion as to whether opioids such as codeine or hydrocodone may accurately be termed "weak" has prompted use of the term "opioid commonly prescribed for moderate pain" in its place. (Although the latter term may appear circular, both are widely used synonymously.) When higher and more frequent doses of opioids are necessary, combination products containing fixed ratios of opioids and NSAIDs should be avoided so as not to exceed inadvertently the maximum recommended dosage of the NSAID. If pain still persists, treatment progresses to the third tier: substitution of the "weak" opioid for a "strong" opioid (i.e., one more readily titrated to doses with greater analgesic efficacy). The latter category includes morphine, hydromorphone, methadone, fentanyl, and levorphanol, all full opioid agonists at the morphine or mu receptor. The WHO approach to managing cancer pain emphasizes by-the-clock rather than as-needed dosing and careful therapy individualized to each patient.

Multiple investigators have reported case series in which the WHO method yields satisfactory pain relief in a majority (80-90%) of patients with cancer pain. However, validation trials of the specific choice of agents and the sequence of their application within the WHO ladder have been limited (Eisenberg, Berkey, Carr, et al., 1994; Jadad and Browman, 1996; Mercadante, 1999). The common clinical impressions that NSAIDs are particularly beneficial for bone pain, or that opioids are of little benefit for neuropathic pain, are either unconfirmed in systematic literature reviews (Eisenberg, Berkey, Carr, et al., 1994) or unsupported by direct clinical trials of mechanism-based drug selection (Ashby, Fleming, Brooksbank, et al., 1992). At present there is no evidence to decide whether treatment strategies based on pain severity (regardless of mechanism) provide outcomes superior to mechanism-based drug selection. Such evidence is likely to emerge slowly because it is difficult to conduct trials in which multiple concurrent pain mechanisms are defined and tracked with precision.

In practice, clinical consensus and common sense dictate initial use of the least invasive delivery method and simplest dosage regimen (Cherny, Chang, Frager, et al., 1995). NSAIDs and certain adjuvants have ceiling effects to analgesic efficacy but not to side effects. NSAID side effects include gastrointestinal distress and bleeding, renal insufficiency or failure, interference with platelet function, and less commonly, allergic reactions (which may impair hepatic function), fluid retention, or central nervous system dysfunction. If pain relief is not achieved at the maximum recommended dose of a particular NSAID or opioid, it should be discontinued and another drug from the same class tried before abandoning that class. Case series indicate that on an individual basis, other drugs from the same class may prove more effective or be better tolerated. Morphine and similar opioids lack a ceiling (or have a much higher ceiling) to their acute analgesic efficacy and so are normally administered in increasing doses until pain relief is obtained or unacceptable side effects occur.

Tolerance and physical dependence are common and to some extent even predictable during chronic opioid administration (Basbaum, 1995). These terms are often confused with psychological dependence ("addiction") that causes drug abuse or drug-seeking behavior. However, tolerance simply refers to the requirement for escalating and/or more frequent doses of an agent in order to sustain therapeutic effectiveness during chronic administration. Physical dependence indicates that, for certain chronically administered drugs (e.g., benzodiazepines or opioids), sudden discontinuation or the presentation of an antagonist drug will precipitate an abstinence syndrome. Misunderstanding these distinctions of nomenclature promotes undertreatment of cancer pain by unnecessarily stigmatizing requests and prescriptions for opioids and their dispensing (Joranson, 1998). Increasing analgesic requirements in a patient with cancer may be a sign of disease progression or recurrence. Patients with nonprogressive disease on a stable regimen usually do not require escalating opioid doses to maintain good pain control. Tolerance should not be confused with "pseudo-addiction," in which the inadequacy of an analgesic dose stimulates sympathetic activity and other signs of acute pain, along with efforts to obtain more medication. Because there is no maximum recommended dose for a full opioid agonist such as morphine, doses of morphine in the range of many grams per day have been administered to some patients. Breakthrough medication, ideally administered in advance of activities such as movement that predictably elicit pain, is added to the by-the-clock regimen.

Oral administration of drugs can manage most cancer pain, but dysphagia from mucositis, nausea from chemotherapy or radiation therapy, malabsorption from gastrointestinal dysfunction such as fistula or dumping syndrome, or the need to swallow an unwieldy number of tablets may indicate the need for other routes of administration. Other noninvasive systemic routes include rectal, transdermal (Breitbart, Chander, Eagel, et al., 2000), sublingual, transmucosal, and pulmonary, as well as subcutaneous drug delivery. Patient-controlled analgesia (PCA) devices have been used to deliver subcutaneous or (when access permits) intravenous or intraspinal medication; oral analgesics may also be given so that patients can self-administer each dose. The subcutaneous route may not be feasible in cachetic patients.

Common side effects that limit opioid dosing are constipation and nausea. Other side effects include sedation, fatigue, vomiting, confusion, urinary retention, pruritus, myoclonus, dysphoria, euphoria, sleep disturbance, sexual dysfunction, respiratory depression, physiologic dependence, tolerance, and endocrinologic abnormalities. Persistent respiratory depression is rare in opioid-tolerant individuals. Addiction rarely occurs in patients with cancer or other medical illness in the absence of a history of substance abuse.

Adjuvant medications are recommended by the WHO at any step of the ladder to provide analgesia for specific types of pain such as neuropathic pain. Adjuvants are also used to treat concurrent symptoms, such as constipation or sedation, or to augment the analgesic efficacy of classes of agents such as opioids. The diversity of drug mechanisms encompassed within the word "adjuvant," and the first-line status of certain agents such as anticonvulsants for neuropathic pain, is not captured by that single word. Adjuvant medications include antidepressants (e.g., tricyclics such as amitriptyline or imipramine), anticonvulsants (e.g., gabapentin or carbamazepine), local anesthetics (e.g., mexilitene), centrally acting muscle relaxants (e.g., baclofen), anxiolytics (e.g., alprazolam, lorazepam, or clonazepam), alpha-2 adrenergic agonists (e.g., clonidine), the N-methyl D-aspartate (NMDA) receptor antagonist ketamine, and corticosteroids.

For the minority of patients whose cancer-related pain is difficult to control with oral opioids, neuraxial drug delivery is available (Bennett, Serafini; Burchiel, et al; 2000; Brown, 1996; Carr and Cousins, 1998). In general this route should be considered only when moderate or severe pain cannot be controlled with systemic drugs because of dose-limiting side effects or toxicity. Considerations as to whether to employ central routes for drug delivery include the site(s), nature, and character of pain; life expectancy; therapeutic preferences of the patient and family; ability of the infrastructure to manage the device and catheter; and stage of the underlying disease (Dougherty and Staats, 1999). A therapeutic trial of neuraxial drug delivery involves percutaneous placement of a temporary epidural or intrathecal catheter in order to gauge the patient's response prior to permanent implantation of a drug delivery device. As described in Chapter 3, controlled clinical trials to examine possible advantages in using neuraxial opioids over systemic opioids for chronic refractory pain are few. Calculations of an effect size or ratio for efficacy of neuraxial versus other drug delivery methods are not possible because nearly every publication is a case series (Filos, Goudas, Patroni, et al., 1993). Outcomes of current combination spinal drug therapies for pain are almost entirely undocumented in controlled trials, of which only a handful examine even single opioid delivery.

Cognitive-behavioral methods seek to help the patient gain or maintain functionality and restore a sense of psychological control (Turk, Sist, Okifuji, et al., 1998). These approaches ordinarily have no negative side effects but do require multidisciplinary, early implementation and active participation by a patient who is able to respond to verbal cues. They also are subject to interference by the cognitive effects of analgesics and other medications. Distraction diverts attention away from pain by the performance of external, enjoyable tasks or by evoking internal images. Recent brain imaging studies disclose alterations in brain function accompanying pain relief produced by distraction. Relaxation is commonly offered to those who experience distressing or escalating levels of pain and anxiety. Cognitive therapy has patients redefine ("reframe") their negative perceptions of bodily sensations including pain into more positive, productive ones. Cognitive therapy is facilitated by simply providing the patient and family with information about pain and the medications employed to treat it. Stress inoculation prepares the patient to employ behavioral pain control methods (e.g., hypnosis and distraction therapy) during an imagined future painful procedure prior to its occurrence. Hypnosis is often used in conjunction with relaxation and cognitive therapy to boost the efficacy of these other approaches. Hypnotic suggestions typically include anesthesia of a painful site, dissociation from one's body so as to perform enjoyable imaginary tasks, and substitution of distressing situations with more pleasant ones.

Nondestructive analgesic approaches generally precede tissue-damaging forms of palliation such as neurolytic blocks and other anesthetic techniques, radiation therapy, or neurosurgical division of afferent pathways. Among the consensus exceptions (see Chapter 3) are celiac block (Brown, Bulley, and Quiel 1987; Brown 1989; Eisenberg, Carr, and Chalmers, 1995) for a patient with pancreatic or other retroperitoneal tumor who presents with moderate to severe pain and palliative radiotherapy for pain at the site of a long bone metastasis or an isolated brain metastasis. The decision to employ neurolytic blocks normally follows inadequate pain control with more conservative therapy, lack of other efficacious options, access to medical and social support systems afterward, and a favorable result from a test block using local anesthetic (Brown 1996; Cousins and Bridenbaugh, 1998). Over one-third of radiation therapy treatments are given for palliation. Curative radiation therapy differs from palliation, in which the goal is to relieve pain quickly while attempting to minimize symptoms. To lessen the burden of travel to and from treatment (an issue especially for patients expected to live for less than 3 months) patients may receive fewer treatments, each of greater dose (McQuay, Carroll, and Moore, 1997). A balance must be struck in relieving the pain caused by metastases (including impending spinal cord compression) and minimizing adverse radiation effects on normal tissue. Radiopharmaceuticals are also used for cancer pain, mainly pain arising from generalized osteoblastic skeletal metastases. Agents such as radiopharmaceuticals are examples of a growing number of clinically available compounds that, while not intended to cure, alter pathophysiology of the underlying disease so as to minimize pain. Chief among these palliative agents are biphosphonates, which are well documented to decrease biochemical and histomorphometric parameters of bone turnover from metastases and also (though not as uniformly) to reduce pain; gemcitabine for pancreatic cancer; and mitoxantrone for prostate cancer.

Issues in the Undertreatment of Cancer Pain

Growing patient and consumer interest in pain control to enhance quality of life during serious illness such as cancer has been paralleled by scientific advances in understanding how pain is generated and sustained (Baszanger, 1998; Chapman and Nakamura, 1999; Wall and Melzack, 1999). These two developments are linked in that societal awareness of the value of effective pain control has prompted public and private support for preclinical research to better understand pain mechanisms and industrial development of novel drugs and delivery systems to meet consumer demand. This same awareness has also led to implementation of clinical standards (Joint Commission, 2000), practice guidelines, and other initiatives to translate these advances into improved patient care. At the state level, an upsurge in pain-related regulations, statutes, and guidelines (see Figure 3) reflects an increased legislative awareness of pain control, in some instances prompted by a concern that patients may seek physician-assisted suicide to escape poorly treated pain and depression (Joranson, 1998).

Figure 3. Approximate number of State pain-related policies in effect, 1980-1998.


Figure 3. Approximate number of State pain-related policies in effect, 1980-1998.
SOURCE: University of Wisconsin Pain Studies Group/WHO Collaborating Center, 1998.

Despite these encouraging trends, and the prospect of achieving cancer pain control in the great majority of patients, cancer pain remains undertreated even in oncology specialty clinics within wealthy, industrialized nations (Cleeland, Gonin, Hatfield, et al 1994). A substantial body of research indicates that this undertreatment is multifactorial. Inadequacy of clinicians' knowledge of effective pain assessment and management, negative attitudes of patients and clinicians toward the use of drugs (particularly opioids) for pain relief (Hill and Fields, 1989; Crothers, 1999 Joranson, Ryan, Gilson, et al., 2000), and problems of access, cost, and reimbursement (Hoffman, 1998) for effective pain management (Bonica, 1990) each contribute. The elderly, women, and members of racial minorities are at increased risk for unsatisfactory pain relief because of poor palliative care practices (Walsh, 2000), patient attitudes (stoicism), and poor communication between patients and providers (Cooper-Smith, Gallo, Gonzales, et al., 1999). In addition, much work is required to pursue clinically relevant pilot studies that document gender (Unruh 1996; Giles and Walker, 1999; Miaskowski and Levine, 1999) and genetic (Gershon, Vatine, Shir, Wu, et al., 2000; Mogil, 1999) differences in effects and side effects of analgesics such as opioids.

Issues of culture and ethnicity (Fadiman, 1997) have considerable importance for cancer pain assessment and management but, like ethnopharmacologic factors, have until recently received little attention in published clinical trials. Driven by changes in immigration, age, and fertility patterns, the ethnic composition of the United States continues to be transformed. Refugees, a substantial proportion of the new immigrants, have been devastated by the impact of civil disorder, trauma, and resettlement (Fadiman, 1997). The design, delivery, and assessment of health care interventions to alleviate their pain therefore presents a complicated and largely unfinished task. The degree to which individuals exhibit behaviors and attitudes of their culture depends upon their educational, occupational, and economic status and other factors such as area of origin within the mother country, religion, acculturation, degree of isolation in their own ethnic social network, and experience with the medical care system (Harwood, 1981). Language barriers that immigrants face exacerbate their economic, social, and health problems. Delivering culturally sensitive care and conducting suitable analgesic trials require knowledge of cultural and linguistic influences on pain expression and description of pain quality, reaction to the pain experience such as seeking health care, coping styles and adopting disability status, the patient/provider relationship, and receptivity to treatment and compliance (Bailey, 1987; Lasch, Wilkes, Leonard, et al., 1999; Moore and Brodsgaard, 1999).

Racial minorities experience a disproportionate share of the cancer burden that includes inadequate pain management (Bernabei, Gambassi, Lapane, et al., 1998; Blendon, Aiken, Freeman, et al., 1989; Cleeland, Gonin, Baez, et al., 1997). Patient, clinician, and systems factors such as insurance status may contribute to the discrepancy in pain treatment between minority and nonminority patients. For example, Calvillo and Flaskerud (1993) found no significant differences between Anglo-American women and Mexican-American women patients' self-report measures of cholecystectomy pain but did find that nurses assigned more pain to Anglo-American patients (Calvillo and Flaskerud, 1993). A recent review of the literature on cross-cultural issues in pain finds both universal and ethnospecific aspects to the experience of pain and its treatment (Moore and Brodsgaard, 1999). The emerging literature on cultural and ethnic influences on pain experience and expression, and analgesia, is of uneven methodological quality. However, there is a clear consensus that for efficacy studies to be valid and clinically useful for patients from diverse cultures, cultural factors must be taken into account (Zatzick and Dimsdale, 1990). Some researchers are content to produce a reliable, valid cross-cultural pain measurement instrument through qualitative groundwork, statistical tools such as multidimensional scaling, and development processes that include translation and back-translation. Others suggest that culture must be taken into account throughout the process, and tools employed that are culturally as well as linguistically appropriate if efficacy studies are to be relevant to specific patient groups. Efficacy studies in cancer pain relief rarely analyze their results for ethnic variation. The emerging demand to provide culturally sensitive care underscores the importance of including ethnic and cultural factors in analgesic trials.


Growing societal attention to the problem of cancer pain may reflect declining mortality in the United States, which has led to an increasing number of older Americans at increased risk to develop cancer. Inroads in reducing the national burden of cancer through early detection and improvements in therapy have not been uniformly successful. Pooled cancer incidence across all ages and races and both sexes has increased during recent decades. The disease burden of cancer remains substantial, particularly in minority groups and the elderly. Since cancer-related pain is more prevalent with advancing stage of disease, the increasing numbers of elderly people in the United States are at increased risk for developing not only cancer but also pain if the disease progresses. Society has recognized these threats to its well-being and has responded in diverse ways. Consumer and legislative involvement in end-of-life issues, including pain management, has never been greater. Journals, textbooks, publications, policy initiatives, and educational efforts such as symposia on cancer-related pain (and the related topics of palliative, supportive, and end-of-life care) have multiplied in recent years. Both the number and rate of increase of high-quality clinical trials in cancer pain control are increasing. However, the total number of relevant trials, and the aggregate number of patients enrolled, is still meager in relation to the numbers of patients affected, and only a small fraction of the published evidence is combinable through quantitative meta-analyses (Abram and Hopgood, 1996; Gray, 1997; Mulrow and Oxman, 1997).

Key conclusions described later in this literature review are that cancer-related pain is quite prevalent, but many treatment modalities are efficacious. Current treatments rely to a great extent on historical approaches, such as morphine- and aspirin-like drugs by mouth, although there have been genuine advances in drug delivery and refinement of agents. Some newer agents for cancer pain control target mechanisms of tumor pathophysiology. Analgesic interventions for treating cancer pain, however, are still for the most part not specific to malignancy. Although authoritative clinical accounts of cancer pain assessment and treatment describe the value of clarifying the mechanism of each pain in each patient, the widely applied WHO ladder stratifies analgesic therapy principally by pain severity rather than by mechanism. A mechanism-based approach to pain therapy is attractive both conceptually and as a first step toward raising pain relief to the level of other medical disciplines that target specific drugs to specific components of complex disorders. However, whether practical application of the latter approach will be more effective than the former is unknown. The optimal group of therapies for pain due to specific cancers, the decision-making process for choosing between them, and their concurrent selection or the sequence in which to apply them in each case are very much open questions.

Substantial impediments to optimal pain management in patients with cancer exist in the form of inconsistent assessment, patient-related barriers (both individual and cultural), provider-related barriers including reluctance to prescribe opioids, regulatory constraints, and other health system barriers such as reimbursement issues (Miaskowski, 1994). Even if many drug and nondrug treatments of proven effectiveness were widely available -- which is close to being the case today -- their mere availability would not by itself cause these other barriers to vanish. At present a unique opportunity exists to advance the quality of the evidence on pain and palliative care by adoption of rigorous research methods (Jadad, 1994; McQuay and Moore, 1998; Cassel and Foley 1999) that have been lagging in these fields (Guyatt, Sinclair, Cook, et al., 1999; National Health and Medical Research Council, 1999).

PubReader format: click here to try


  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...