Investigations of the epidemiology of prescription opioid addiction in CNCP have provided disparate estimates of the rates of this potential complication of chronic opioid therapy. A letter reporting addiction in only 4 hospitalized patients out of 11,882 treated with narcotics has been cited in support of low addiction rates in opioid-treated patients (Porter and Jick 1980). However, results observed in hospitalized patients may not generalize to an outpatient population; furthermore, this study documented opioid addiction only in patients with no prior addiction history, thereby potentially missing many addiction cases. An early study by Portenoy and Foley reported a 5% rate of addiction in CNCP patients (Portenoy and Foley 1986), while a prospective study of patients referred to an interventional pain clinic found the rate of opioid abuse to be 9% (Manchikanti et al. 2006). An early systematic review reported addiction rates of 3.2–18.9% in CNCP patients (Fishbain et al. 1992). A higher rate of opioid misuse (32%) was observed in a study of patients referred to a Veterans Administration (VA) primary pain clinic, with predictive factors being prior history of addiction or conviction on drug charges (Ives et al. 2006). Still higher levels of problematic opioid use among CNCP patients were observed in two additional studies, which together reported rates of opioid misuse or dependence between 20 and 40% (Reid et al. 2002; Katz and Fanciullo 2002). The varied rates of opioid use disorders cited above likely reflect not only differences in study populations, but also differences in definitions of abuse, addiction and misuse, and in outcome measures used. Studies variously assessed opioid misuse, abuse, and dependence. The choice to measure different outcomes in these patient populations may in part reflect the difficulties and ambiguities in making opioid use diagnoses in CNCP patients. Clearly, there is a need for additional research to clarify the discrepancies in the epidemiology of opioid use disorders in CNCP patients.

Clinicians have identified certain ADRB that are associated with opioid dependence (Fleming et al. 2008). Among these, it has been suggested that certain behaviors (forging prescriptions, stealing or borrowing drugs, frequently “losing” prescriptions, and resisting medication changes) are more predictive of opioid misuse than other behaviors (aggressive complaining about the need for more drugs, drug hoarding during periods of reduced symptoms, and unsanctioned dose escalations) as the latter may indicate poorly controlled pain (Portenoy 1996). There is some overlap between ADRB and DSM-IV criteria for opioid dependence (e.g., self-escalation in opioid dose may be analogous to more use than intended in DSM); however, not all patients with ADRB will meet criteria for DSM-IV opioid abuse or dependence. Thus, the presence of ADRB may alert clinicians to the presence of an opioid use disorder or may indicate vulnerability for developing an opioid use disorder.

Several scales and diagnostic tools have been developed to assess the presence of ADRB in patients receiving opioid analgesics. For example, the revised version of the Screener and Opioid Assessment for Patients with Pain (SOAPP-R) is a self-report questionnaire that includes questions about family and personal history of substance abuse problems, lost or stolen medications, legal problems, and medication craving (Butler et al. 2008). The same group developed another questionnaire, the Current Opioid Misuse Measure (COMM), which was designed to monitor the misuse of medication in patients who have been prescribed opioids for extended time periods (Butler et al. 2007). Another group developed a different questionnaire, the Opioid Risk Tool (ORT), which was developed to predict which patients will develop ADRB (Webster and Webster 2005). The ORT includes questions about family and personal history of substance abuse as well as questions about psychiatric diagnoses and history of sexual abuse. The Prescription Drug Use Questionnaire (PDUQ) is an interview screening tool created to stratify patients into three groups: those who are likely to be (1) non-addicted, (2) substance-abusing, and (3) substance-dependent (Compton et al. 1998). A limitation of the PDUQ is that it is designed to be used by a trained mental health professional, and thus may be difficult to incorporate into a primary care setting. Finally, the Diagnosis, Intractibility, Risk, and Efficacy (DIRE) Score is another clinician-scored tool that relies on chart review to assign scores based on diagnosis, engagement in treatment, history of psychiatric disorders and substance abuse, and functional status (Belgrade et al. 2006). The DIRE Score was developed with the aim of predicting compliance and efficacy with the ultimate goal of helping clinicians predict which patients would benefit most from opioid therapy.

Efforts have been made to predict which CNCP patients are most at risk for prescription opioid addiction when prescribed opioid analgesics. Although these predictive factors are currently imperfect, certain risk factors for prescription opioid addiction have been identified. As might be expected, one of the most consistent predictors of prescription opioid misuse is a prior history of substance use disorders (Ives et al. 2006; Edlund et al. 2007; Turk et al. 2008). For example, a study conducted at a VA multidisciplinary opioid renewal clinic (ORC), a history of cocaine abuse was the most powerful predictor for failing the program, increasing the odds of program failure by approximately 5 times (odds ratio 4.97) (Meghani et al. 2009). In addition, tobacco smokers may be at greater risk for opioid misuse (Michna et al. 2004). A family history of substance abuse may also be predictive of opioid misuse (Schieffer et al. 2005). The clinical utility of self-reported history of substance use may be limited by patient reluctance to divulge these behaviors in themselves and in their significant others (Cook et al. 1995; Fishbain et al. 1999). Urine drug screens (UDS) can potentially address problems associated with relying on self-report, although they cannot provide diagnoses of substance abuse or dependence. A history of legal problems has also been observed to predict opioid misuse. For example, in one study a history of DUI or drug convictions were strong predictors of opioid misuse (Ives et al. 2006). Younger age may also predict greater risk for prescription opioid misuse (Turk et al. 2008).

The presence of ADRB can alert clinicians about the potential for opioid addiction. Fleming et al. (2008) showed that a significantly greater proportion of patients with four or more lifetime ADRBs (vs. those with one to three lifetime ADRBs) met DSM-IV criteria for opioid dependence. Even so, most patients who engage in ADRB do not appear to have opioid dependence, as this same study also reported that only 9.9% of patients with four or more lifetime ADRB met criteria for opioid dependence (Fleming et al. 2008). Furthermore, in a study of patients referred because of the presence of ADRB [performed at the PVAMC Opioid Renewal Clinic (ORC)], only 13% required referral for addiction treatment when provided a structured approach to opioid management that included opioid treatment agreements and UDSs (Wiedemer et al. 2007). It is possible that other patients with ADRB may have pseudoaddiction, or they engage in ADRB due to undertreatment of pain (Gallagher and Rosenthal 2008). It is also possible that certain ADRB are more predictive of substance use problems than others. For example, Fleming and colleagues propose that four specific behaviors appear useful as screening questions to predict patients at risk for a current substance use disorder based on the high proportion of positive responses in this group compared to low risk patients: (1) oversedated oneself, (2) felt intoxicated, (3) early refills, and (4) increased dose on own (Fleming et al. 2008). Another group reported that craving for an opioid medication predicted a higher incidence of physician-rated ADRB and more frequent positive UDS (Wasan et al. 2009).

As with all medications, clinicians must weigh the risks and benefits of opioid analgesic therapy for each individual patient. Balancing risks and benefits is particularly challenging in the case of opioid analgesics (Gallagher and Rosenthal 2008; Ballantyne 2007), as clinicians must not only consider benefits and traditional side effects/serious adverse events, but also the risk of addiction, which patients may not always recognize or divulge. These difficulties are compounded by lack of clarity in the terminology and diagnosis related to prescription opioid addiction, which results in a wide “gray area” between patients who clearly benefit from chronic opioid therapy vs. those who are clearly harmed.

Opioid therapy can be a useful treatment strategy to improve pain symptoms and quality of life in some patients with CNCP. The risks and benefits of opioid therapy must be carefully weighed for each individual patient, and a trial of opioid analgesics begun in those patients for whom the likely benefits outweigh potential risks. This chapter presents the factors that clinicians must weigh in the decision to begin opioid therapy for CNCP conditions. In addition, we present some practical strategies for addressing problems that may arise in the course of opioid therapy. Though in some cases there is insufficient evidence for definitive guidance of clinical decisions, it is hoped that these questions will be addressed in future research.

A number of issues show promise as potential areas of future research to improve the use of opioid analgesics to treat CNCP. For example, little is known about the effects of opioids on the brains of CNCP patients, or indeed about the long-term effects of opioids on the brain in general. Thus, more neuroimaging studies in these areas would be helpful. As mentioned above, there is substantial individual variability in response to opioids. In addition, there may be variability in response to opioids among different CNCP conditions. Predicting which patients are more likely to respond to opioid analgesics, and furthermore which patients will respond to specific opioids would help clinicians better guide treatment choices. Promising areas of future research also include the search for novel therapeutic compounds. For example, in 1995 a novel receptor with similarities to the known opioid receptors was identified (ORL-1, or “opioid-receptor-like 1” Chiou et al. 2007; Reinscheid et al. 1995; Meunier et al. 1995). Although there are no currently available medications that specifically target this receptor, it is possible that the ORL-1 receptor may become the target of novel analgesic medications.