|NCI/PDQ® Health professionals: Substance Abuse Issues In Cancer (PDQ®)|
|National Cancer Institute|
| Last Modified: June 30, 2011
Substance abuse in cancer patients who do not have histories of substance abuse is exceptionally rare. Opioids and other controlled substances can be prescribed judiciously for symptom management, without concern about misuse. When problematic drug-taking behavior is manifested by such patients, it is often the result of poor pain control. However, many people with cancer have a history of drug abuse or live among those who do. They have special needs that are often underappreciated because this problem is overlooked. 1 Nearly one-third of the population of the United States has used illicit drugs, and an estimated 6% to 27% have a substance abuse problem of some type. 2 3 4 5 The abuse of prescription opioids has grown rapidly since the mid-1980s and is now as frequent as the abuse of cocaine. 6 Because of the prevalence of substance abuse and the association between drug abuse and some types of cancer, 1 problems related to abuse and addictions are encountered in palliative care settings.
The population of patients who have histories of substance abuse or addiction is extremely heterogeneous, and the status of each patient will affect concerns central to his or her palliative care. Patients who are actively abusing alcohol, illicit drugs, or prescription drugs present problems distinct from those experienced by patients in drug-free recovery or patients in methadone maintenance programs. Appropriate diagnosis of substance abuse may be challenging because of the variability in abuse behaviors over time, the changes in comorbid physical and psychosocial factors that influence drug abuse, and the problems inherent in the nomenclature of drug abuse in the physically ill.
Patients who have histories of substance abuse present many clinical problems. Clinicians must control and monitor drug use in all patients. Compliance with treatments for the underlying disease may be so poor among cancer patients who are actively abusing drugs that the substance abuse actually shortens life expectancy by preventing the effective administration of primary therapy. Prognosis may also be altered by the use of drugs in a manner that negatively interacts with therapy or predisposes the patient to other serious morbidity.
Active or past substance abuse also may weaken social support networks. Among these supports is the patient's relationship with the treatment team. Lack of mutual trust can characterize the relationships between substance-abusing patients and members of the treatment team. Concerns about drug abuse may lead clinicians to doubt the veracity of the history divulged by the patient, the report of symptoms, and compliance with therapy. A desire to build trust may lead clinicians to hide these concerns from the patient. Patients with histories of substance abuse may sense the mistrust, question the team's good will, and have negative expectations that become self-fulfilling prophecies. Mistrust can disrupt assessment, management, and follow-up and can result in the failure of therapies intended to improve quality of life.
Thus, a history of substance abuse can undermine palliative care and increase the risk of morbidity or mortality among patients with progressive, life-threatening diseases. This potential can only be mitigated by a therapeutic strategy that addresses drug-taking behavior while implementing other therapies. To organize this strategy, clinicians who provide palliative care in the cancer setting must be knowledgeable about the basic concepts of addiction medicine.
In this summary, unless otherwise stated, evidence and practice issues as they relate to adults are discussed. The evidence and application to practice related to children may differ significantly from information related to adults. When specific information about the care of children is available, it is summarized under its own heading.
Substance abuse appears to be very uncommon among cancer patients. The reported prevalence of substance abuse issues in cancer patients is much lower than the prevalence in society at large, general medical populations, and emergency medical departments. 1 2 3 4 5 This relatively low prevalence was also reported in a Psychiatric Collaborative Oncology Group study, which assessed psychiatric diagnoses in ambulatory cancer patients from several tertiary care hospitals. 5 On the basis of structured clinical interviews, fewer than 5% of 215 cancer patients met the Diagnostic and Statistical Manual for Mental Disorders, 3rd Edition (DSM-III) criteria for a substance use disorder. 6
The relatively low prevalence of substance abuse among cancer patients treated in tertiary care hospitals may reflect institutional biases or a tendency for patient underreporting in these settings. Drug abusers who are poor or feel alienated from the health care system may not seek care in tertiary centers. Those who are treated in these centers may not acknowledge a history of drug abuse. The low prevalence of drug abuse in cancer centers, therefore, may not represent the true prevalence among the cancer population overall. A survey of patients who were admitted to a palliative care unit found indications of alcohol abuse in more than 25% of these patients; 7 this, however, represents highly selected palliative care patients referred to a specialized inpatient unit.
Epidemiologic studies and clinical management depend on an accepted, valid nomenclature for substance abuse and addiction. The pharmacologic phenomena of tolerance and physical dependence are commonly confused with abuse and addiction. Terminology is also strongly influenced by sociocultural considerations that may lead to the sending of mixed messages in the clinical setting. The definitions of addiction and abuse that are applied to patients who are physically ill have been developed from populations of addicts who do not have physical illnesses. The clarification of this terminology is an essential step in improving the diagnosis and management of substance abuse in the palliative care setting. The list below gives the proposed definitions for these terms.
Physical dependence is defined solely by the occurrence of a characteristic withdrawal or abstinence syndrome following discontinuation of dosing, abrupt dose reduction, or administration of a pharmacologic antagonist. 8 9 10 Neither the dose nor duration of administration of opioids required to produce clinically significant physical dependence in humans is known. Most practitioners assume that the potential for abstinence syndrome exists after opioids have been administered repeatedly for only a few days.
Physical dependence is not apparent unless abstinence is induced. In the clinical setting, physical dependence on an opioid is not considered a problem as long as patients are told to avoid abrupt discontinuation of therapy and to avoid the inadvertent administration of an opioid antagonist (including an analgesic from the agonist-antagonist class).
There is often confusion among clinicians about the differences between physical dependence and addiction. Physical dependence, like tolerance, has been suggested to be a component of addiction, 11 12 and the avoidance of withdrawal has been postulated to create behavioral contingencies that reinforce drug-seeking behavior. 13 These speculations, however, are not supported by experience with opioid therapy for chronic pain. Physical dependence does not preclude the uncomplicated discontinuation of opioids during multidisciplinary pain management of nonmalignant pain. 14 Opioid therapy is routinely stopped without difficulty in cancer patients whose pain disappears following effective antineoplastic therapy. Indirect evidence for a fundamental distinction between physical dependence and addiction is provided by animal models of opioid self-administration, which have demonstrated that persistent drug-taking behavior can be maintained in the absence of physical dependence. 15
Tolerance, a pharmacologic property defined by the need for increasing doses to maintain effects, 8 9 has been a particular concern with opioid therapy. Clinicians and patients commonly express concern that tolerance to the analgesic effects of opioids may compromise the benefits of therapy and lead to a requirement for progressively higher and ultimately unsustainable doses. It has been speculated that the development of tolerance to the reinforcing effects of opioids and the consequent need to increase doses to regain these effects is an important element in the pathogenesis of addiction. 13
Despite these concerns, extensive clinical experience with opioid drugs given for medical reasons has not confirmed that tolerance causes substantial problems. 16 17 Numerous surveys have demonstrated that most patients can attain stable opioid doses with a favorable balance between analgesia and side effects for prolonged periods.
Clinical observation does not support the conclusion that analgesic tolerance is a substantial contributor to the development of addiction. It is widely accepted that addicts who do not have a medical disorder may not have any of the manifestations of analgesic tolerance. Opioid-treated patients who present with analgesic tolerance typically do so without evidence of abuse or addiction.
The definitions of tolerance and physical dependence highlight deficiencies in the current nomenclature applied to substance abuse. The terms addiction and addict are particularly troublesome. In common parlance, these labels are often inappropriately applied to describe both aberrant drug use (reminiscent of the behaviors that characterize active abusers of illicit drugs) and phenomena related to tolerance or physical dependence. Clinicians and patients may use the word addicted to describe compulsive drug-taking in one patient and to describe nothing more than the possibility for withdrawal symptoms in another. It is not surprising that patients, families, and staff become very concerned about the outcome of opioid treatment when this term is applied.
The labels addict and addiction should never be used to describe patients who are only perceived to have the capacity for abstinence syndrome. These patients must be labeled physically dependent. Use of the word dependent alone also should be discouraged because it fosters confusion between physical dependence and psychological dependence, a component of addiction. For the same reason, the term habituation should not be used.
The psychiatric terminology applied to drug abuse and addiction, which has been codified in the DSM-IV, is also problematic. 12 The DSM-IV eschews the term addiction altogether and offers definitions of two types of substance use disorders: substance abuse and the more serious substance dependence. The criteria for substance abuse are focused on the negative psychosocial sequelae of drug use rather than the pattern of use. In contrast, a pattern of use outside of sociocultural convention is considered to be the most important criterion for abuse in other definitions (see list on Proposed Terminology for Substance Abuse). 18 19 The disparity in these definitions of abuse is confusing and underscores the challenge in labeling drug-taking behaviors in patients who are receiving potentially abusable drugs for legitimate medical purposes.
The DSM-IV criteria for substance dependence highlight psychosocial dysfunction and add the dimensions of physical dependence and tolerance. This is perhaps the most striking example of the nomenclatural problems that occur when criteria developed in substance abusers without medical illnesses are applied in a different context. 20 Most of the criteria for substance dependence disorder indicate that the term is meant to be used in a manner synonymous with addiction. The criteria of tolerance and physical dependence, therefore, are inappropriate and preclude the use of this terminology in the medically ill who may develop tolerance and physical dependence as expected consequences of therapeutic drug use.
Undertreatment, sociocultural influence on the definition of aberrancy, and disease-related variables are concerns that increase the difficulty of assessing drug-taking behavior.
Clinical observation suggests that inadequate management of symptoms may be an impetus for aberrant drug-related behaviors. This concept has been extensively explored in the area of cancer pain. There is compelling evidence that pain is undertreated in populations of medically ill patients, including those with cancer and AIDS. 1 2 The term pseudoaddiction was coined to depict the distress and drug-seeking that can occur in the context of unrelieved cancer pain. 3 The cardinal feature of this syndrome is that the aberrant behaviors disappear when an effective analgesic intervention is administered. In the cancer population, first-line intervention is often a higher dose of an opioid.
Assessment for pseudoaddiction in the population of known substance abusers who develop painful medical disease is a challenge for clinicians. Clinical experience suggests that aberrant behaviors driven by unrelieved pain can become dramatic or particularly worrisome in substance abusers. Some patients appear to return to illicit drug use as a means of self-medication, at least in part. Others adopt patterns of behavior with health care providers that also generate intense concern about the possibility of true addiction. Although it may be clear that the drug-related behaviors are aberrant, the meaning of these behaviors may be difficult to discern in the context of unrelieved symptoms. Management strategies must reflect the diagnostic complexity.
When a drug is prescribed for a legitimate medical purpose, there is decreasing certainty about the behaviors that could be characterized as aberrant, abusive, or addictive. Although the aberrancy of some behaviors would not be argued (e.g., prescription forgery or the intravenous injection of an oral formulation), many other behaviors are less clear-cut. For example, is it aberrant for the patient with unrelieved pain to consume extra doses of a prescribed opioid, particularly if this behavior was not specifically prescribed by the clinician? Is it aberrant to use an opioid drug prescribed for pain as a nighttime hypnotic?
The importance of social and cultural norms raises the possibility of bias in determinations of aberrancy. Bias against a social group, even if subtle, could influence the willingness of clinicians to label a questionable drug-related behavior as aberrant when performed by a member of that group. Clinical observation suggests that this type of bias is common in the assessment of drug-related behaviors of patients with substance abuse histories. Questionable behaviors by such patients may be promptly labeled as abuse or addiction, even if the drug abuse history was in the remote past. In a similar way, the possibility of bias in the assessment of drug-related behaviors exists for patients who are members of racial or ethnic groups different from that of the clinician.
The core concepts used to define addiction may also be complicated by changes resulting from progressive disease. Deterioration in physical or psychosocial functioning that is caused by the disease and its treatment may be difficult to separate from the morbidity caused by substance abuse. This may particularly complicate efforts to evaluate the concept of use despite harm, which is critical to the diagnosis of addiction. For example, the nature of questionable drug-related behaviors can be difficult to discern in the patient who develops social withdrawal or cognitive changes following brain irradiation for metastases. Even if impaired cognition is clearly related to the drugs used to treat symptoms, this outcome might reflect a low therapeutic index rather than the patient's desire for these psychic effects.
The accurate assessment of drug-related behaviors in patients with advanced medical disease usually requires detailed information about the role of the drug in the patient's life. Time spent out of bed or the existence of mild mental clouding may be less meaningful than other outcomes, such as noncompliance with primary therapy because of drug use or behaviors that jeopardize relationships with physicians, other health care providers, or family members.
The foregoing discussion emphasizes the difficulties inherent in formulating and applying a nomenclature that would allow appropriate diagnosis of drug-related phenomena in the medically ill. Previous definitions that include phenomena related to physical dependence or tolerance cannot be the model terminology for medically ill populations who receive potentially abusable drugs for legitimate medical purposes. A more appropriate model definition of addiction notes that it is a chronic disorder characterized by the compulsive use of a substance resulting in physical, psychological or social harm to the user and continued use despite that harm. 4 Although this definition was developed from experience in populations of addicts who do not have medical illness, it appropriately emphasizes that addiction is, fundamentally, a psychological and behavioral syndrome. Any appropriate definition of addiction must include several important characteristics, including loss of control over drug use, compulsive drug use, and continued use despite harm.
Even appropriate definitions will have limited utility unless operationalized for a clinical setting. The concept of aberrant drug-related behavior is a useful first step in operationalizing the definitions of abuse and addiction. The concept also recognizes the broad range of behaviors that may be considered problematic by prescribers. Although the assessment and interpretation of these behaviors can be challenging, the occurrence of aberrant behavior signals the need to re-evaluate and manage drug-taking, even in the context of an appropriate medical indication for a drug.
In assessing the differential diagnosis for drug-related behavior, it is useful to consider the degree of aberrancy. Less aberrant behaviors (such as aggressively complaining about the need for medications) are more likely to reflect untreated distress of some type, rather than addiction-related concerns. Conversely, more aberrant behaviors (such as injection of an oral formulation) are more likely to reflect true addiction. Although empirical studies are needed to validate this conceptualization, it may be a useful model when evaluating aberrant behaviors.
Extensive worldwide experience in the long-term management of cancer pain with opioid drugs has demonstrated that opioid administration in cancer patients with no histories of substance abuse is only rarely associated with the development of significant abuse or addiction. 1 2 3 4 5 6 7 8 9 10 11 Indeed, concerns about addiction in this population are now characterized by an interesting paradox. Although the lay public and inexperienced clinicians still fear the development of addiction when opioids are used to treat cancer pain, specialists in cancer pain and palliative care widely believe that the major problem related to addiction is the persistent undertreatment of pain driven by inappropriate fear of addiction.
The experience in the cancer population has contributed to a desire for a reappraisal of the risks and benefits associated with the long-term opioid treatment of chronic nonmalignant pain. 12 13 The traditional view of this therapy is negative. Early surveys, which noted that a relatively large proportion of addicts began their addiction as medical patients who received opioid drugs for pain, provided some indirect support for this perspective. 14 15 16 The most influential of these surveys recorded a history of medical opioid use for pain in 27% of white male addicts and 1.2% of black male addicts. 16
Surveys of addict populations, however, do not provide a valid measure of the addiction susceptibility associated with chronic opioid therapy in populations without known abuse histories. Prospective patient surveys are needed to define this risk accurately. The Boston Collaborative Drug Surveillance Project evaluated 11,882 inpatients who had no histories of addiction and were administered an opioid while hospitalized; only four cases of addiction could be identified subsequently. 17 A national survey of burn centers could find no cases of addiction in a sample of more than 10,000 patients without histories of drug abuse who were administered opioids for pain, 18 and a survey of a large headache clinic identified opioid abuse in only 3 of 2,369 patients admitted for treatment, most of whom had access to opioids. 19
Other data suggest that the typical patient with chronic pain differs significantly enough from the addict without painful disease that the risk of addiction during therapy for pain is low. For example, surveys of cancer patients and postoperative patients indicate that euphoria (a phenomenon believed to be common during the abuse of opioids) is extremely uncommon following administration of an opioid for pain; dysphoria is observed more typically in those who receive meperidine. 20 Although the psychiatric comorbidity identified in addict populations could be an effect, rather than a cause, of the aberrant drug taking, the association suggests the existence of psychologic risk factors for addiction. The likelihood of genetically determined risk factors for addiction has also been suggested by a twin study that demonstrated a significant concordance rate for aberrant drug-related behaviors. 21
Overall, the evidence generally supports the idea that opioid therapy in patients with chronic pain and no histories of abuse or addiction can be undertaken with a very low risk of these adverse outcomes. This is particularly so in the older patient, who has had ample time to reveal a propensity for abuse. There is no substantive support that large numbers of individuals with no personal or family histories of abuse or addiction, no affiliations with substance-abusing subcultures, and no significant premorbid psychopathologies will develop abuse or addiction when administered potentially abusable drugs for medical indications.
The inaccurate perception that opioid therapy always has a high likelihood of addiction has encouraged assumptions that are not supportable in populations with no histories of substance abuse. For example, agonist-antagonist opioid analgesics are less likely to be abused by addicts than pure mu agonist opioids. Consequently, some clinicians view the agonist-antagonist drugs as safer in terms of addiction liability. There is no evidence for this conclusion in populations without drug abuse histories. Extensive experience with long-term opioid therapy for cancer pain and chronic nonmalignant pain 13 22 23 24 25 26 27 has relied on pure mu agonists. Similarly, there is a common perception that short-acting oral opioids and opioids delivered by the parenteral route carry a greater risk of addiction because of the rapid delivery of the drug. Again, these perceptions are derived from observations in the healthy addict population and are not relevant to the treatment of pain in medical patients with no histories of substance abuse.
There is little information about the risk of abuse or addiction during or after the therapeutic administration of a potentially abusable drug to patients with current or remote histories of abuse or addiction. Anecdotal reports have suggested that successful long-term opioid therapy in patients with cancer pain or chronic nonmalignant pain is possible, particularly if the history of abuse or addiction is remote. 1 2 Although there is a lack of empirical information on this subject, it is generally accepted that the risk of aberrant drug-related behaviors during treatment for a medical disorder is higher among populations with current histories of substance abuse and, to a lesser extent, remote histories.
The most prudent actions cannot obviate risk, and clinicians must recognize that virtually any drug that acts on the central nervous system can be abused through any route of drug administration. Effective risk management of patients with substance abuse histories necessitates a comprehensive approach that recognizes the biologic, chemical, social, and psychiatric aspects of substance abuse and addiction, as well as a practical means to manage risk.
The population of patients with substance abuse histories is extremely heterogeneous. The most difficult issues in palliative care typically present in those who are actively abusing alcohol or other drugs. Although the principles in this section can also apply to patients who are in drug-free recovery and those who are in methadone maintenance programs, they are likely to be most helpful in the treatment of the active drug abuser.
The clinical assessment of drug-taking behaviors in medically ill patients with pain is complex. Aberrant drug-taking behavior in cancer pain management is generally related to premorbid history of drug addiction and the likelihood of other pain treatment. A pilot questionnaire was used to characterize drug-related behaviors and attitudes in cancer and AIDS patients. Despite limitations, this study highlights wide potential variation in different patient populations in patterns of past and present aberrant drug-taking behaviors and the need for a clinically useful screening approach. 1[Level of evidence: II] The implications for psychosocial and pharmacological management of symptoms such as pain, as well as any aberrant behavior, remain unclear.
Recommendations for the long-term administration of potentially abusable drugs such as the opioids to patients with histories of substance abuse are based solely on clinical experience. Studies are needed to determine the most effective therapeutic strategies and to empirically define patient subgroups that may be most amenable to different approaches. The following guidelines broadly reflect the types of interventions that might be considered in this clinical context.
In the population of patients with progressive medical illness and substance abuse, palliative care often must contend with multiple medical, psychosocial, and administrative problems. A team approach can be very useful in addressing these problems. The most knowledgeable team may involve one or more physicians with expertise in palliative care, nurses, social workers, and, if possible, one or more mental health care providers with expertise in addiction medicine. 2 3 4[Level of evidence: III]
Drug abuse and addiction often remit and relapse. The risk of relapse is likely to be enhanced because of the heightened stress associated with life-threatening disease and the ready availability of centrally acting drugs prescribed for symptom control. Preventing relapses may be impossible in such a setting. Conflict with staff may be lessened if there is a general understanding that unerring compliance is not a realistic goal of management. Rather, the goal might be the creation of a structure for therapy that includes sufficient support and limit-setting to contain the harm done by occasional relapses.
A small subgroup of patients may be incapable of complying with the requirements of therapy because of severe substance abuse and associated psychiatric comorbidities. To establish the intractability of the problem, clinicians must re-establish limits and attempt to develop an increasing variety and intensity of supports. Frequent team meetings and consultations with other clinicians who have expertise in palliative care and addiction medicine may be needed. Ultimately, appropriate expectations must be clarified, and therapy that is failing cannot be continued in the same way. The success rate for converting highly problematic therapies into those that can be managed over time is unknown.
The comorbidity of depression, anxiety, and personality disorders in alcoholics and other patients with substance abuse histories is extremely high. 5The treatment of anxiety and depression can increase patient comfort and possibly diminish the likelihood of relapse.
Clinicians must be familiar with the signs and symptoms associated with abstinence from opioids and other drugs. Many patients with histories of drug abuse consume multiple drugs. A complete drug use history must be elicited to prepare for the possibility of withdrawal. Delayed abstinence syndromes, such as may occur following abuse of some benzodiazepine drugs, may pose a particular diagnostic challenge.
Patients who are actively abusing drugs may have sufficient tolerance to influence the use of prescription drugs subsequently administered for an appropriate medical indication. Tolerance is a complex phenomenon, and its impact on clinical management in this context is likely to be highly variable. 6 7 A prospective open-label study compared morphine dosage and effectiveness in AIDS patients with and without histories of substance use. Results demonstrated that both groups benefited, but patients with histories of drug use required higher morphine doses to achieve stable pain control. 8 This study should increase confidence in providing patients with histories of drug use with appropriate pain management.
Individualization of the dose without regard to its size is the most important guideline for long-term opioid therapy and can be problematic in patients with histories of substance abuse. Although it may be appropriate to exercise caution in prescribing potentially abusable drugs to this population, the decision to forego the principle of dose individualization without regard to absolute dose may increase the likelihood of undertreatment. 9 10[Level of evidence: II] The resulting unrelieved pain can lead to the development of aberrant drug-related behaviors. Although these behaviors might be best understood as pseudoaddiction, their occurrence confirms clinicians' fears and encourages even greater caution in prescribing.
This cycle must be recognized and openly acknowledged to the patient and the staff. The request for dose escalation should not by itself be viewed as aberrant drug-related behavior, but the concerns it generates should be discussed. If the clinician perceives that limits on prescribing are necessary to assess or manage a problematic therapy, frequent monitoring and alternative approaches to pain control might be offered. The patient should be given clear guidelines for responsible drug-taking with the expectation that responsible drug-taking on the part of the patient will reassure the physician that dose escalation is appropriate.
All patients who are prescribed potentially abusable drugs must be carefully monitored over time for the development of aberrant drug-related behaviors. The need for this monitoring is especially strong when patients have a remote or current history of substance abuse, including alcohol abuse. If there is a high level of concern about such behaviors, monitoring may require relatively frequent visits and regular assessment by significant others who can provide observations about the patient's drug use.
To facilitate the early recognition of aberrant drug-related behaviors in patients who have been actively abusing drugs in the recent past, regular screening of urine for illicit or licit-but-unprescribed drugs may be appropriate. The patient should be informed about this approach, which should be explained as a method of monitoring that can be reassuring to the clinician and can provide a foundation for aggressive symptom-oriented treatments. Presented in this manner, screening is a technique that enhances a therapeutic alliance with the patient. Patients who protest excessively may be unwilling, or unable, to enter a collaborative relationship with the clinician in which the clinician can be confident of responsible drug-taking by the patient. Similarly, the patient can be confident that the clinician will respond to unrelieve