Subscription is FREE for qualified healthcare professionals in the US.
7 Articles in Volume 5, Issue #1
Diagnosing and Managing Interstitial Cystitis
Intractable Pain Centers’ Treatment Approach
Musculo-Skeletal Diagnostic Ultrasound Imaging
Pain Management Pitfalls
Selection Criteria for Intrathecal Opioid Therapy: A Re-examination of the "Science"
‘High Dosage’ Opioid Management
‘Opiophobia’ Past and Present

Selection Criteria for Intrathecal Opioid Therapy: A Re-examination of the "Science"

A survey of the "science" behind exclusionary psychological assessment finds that supporting empirical data is virtually non-existent and may be unduly limiting access to those patients for whom this treatment represents their only chance for comfort and a reasonable quality of life.

Since the pioneering work of Wang et al.1 twenty-five years ago, intrathecal opioid delivery has been used to manage a wide variety of refractory pain conditions in patients whose responses to oral opioid analgesics has been unsatisfactory. While most chronic pain sufferers who require chronic opioid therapy can be managed with oral agents, problems with intolerable side effects and habituation preclude their long-term utilization among a subgroup of chronic pain patients. Coombs et al.2 reported the first successful surgical implantation of an intrathecal opioid pump in 1982. Subsequently, chronic pain conditions which have been found to be treated effectively through intrathecally-administered opioids have included nonmalignant conditions such as failed spinal surgery syndrome, spinal or radicular pain without spinal surgery, complex regional pain syndrome, osteoporotic crush fractures and chronic pancreatitis,3 multiple sclerosis and compression fractures,4 phantom limb pain, arachnoiditis and spinal cord injury,5 loin pain hematuria syndrome,6 severe burns,7 and rheumatoid arthritis8 as well as a number of types of cancer pain.4,9-13

The numerous studies which have supported the use of intrathecal opioid therapy (IOT) for cancer pain have served to make this type of treatment for cancer patients progressively less controversial over the years. However, problems with side effects and risk issues along with an inadequate understanding of the potential improvement in the quality of life of patients suffering from severe intractable pain from nonmalignant causes has perpetuated the controversy regarding the use of IOT among patients with chronic pain of benign cause. This controversy is curious given the lack of research findings suggesting that IOT is not appropriate for patients with intractable nonmalignant pain, particularly when so many investigators3-8,14,15 have determined that IOT can be so effective for these patients, with benefits clearly outweighing potential problems. A review of the literature suggests that only the results of a study by Yoshida and colleagues16 would suggest the general contraindication of IOT for nonmalignant chronic pain. It should be noted, however, that their conclusion was based upon complete data from only 14 patients, thereby impacting its reliability. IOT has also been found to be cost-effective when compared to conventional pain therapies among nonmalignant chronic pain sufferers.8,15,17,18

Concerns Associated With IOT

Despite the wealth of studies supporting the use of IOT with malignant and nonmalignant chronic pain patients, it would be a mistake to suggest that this method of delivering opioid analgesics is completely free of complications. Problems with the catheter, including kinking, obstruction, dislodgement and disconnection, are relatively common, reportedly occurring in 10% to 40% of cases.19 Cerebrospinal fluid leaks are typically evidenced during intrathecal catheter placement, with postspinal headaches reported in up to 31% of patients.20 While these headaches usually resolve spontaneously, some will require an autologous blood patch. Infection can be a problem associated with the implantation of indwelling opioid pumps, as is true for all surgical procedures. Deer et al.14 recently reported an infection rate of 2.2% among patients in a large multi-site study of patients receiving IOT. Aggressive prophylactic use of antibiotics, along with strict sterile technique, has been recommended as a means of reducing infection.19 Despite precautions and improvements in equipment and techniques, the need for additional surgery following the implantation of an intrathecal opioid pump was reported to be 40% in a 2001 Australian study.3 This figure approximated the percentage of patients requiring additional surgery due to complications reported by Tutak and Doleys.21

The accepted advantage of IOT over systemic opioids is the delivery of the opioid directly to the effective site of action in the spinal cord, thereby avoiding brain sites that produce the most troublesome side effects of oral opioids.22 Chronic oral opioid use has been associated with problems with working memory, vigilance/attention, psychomotor speed, and recognition.23,24 Nevertheless, reports in the literature indicate that severe side effects can be associated with long-term IOT. Penn22 noted that the same adverse effects of systemic opioids can occur in patients receiving IOT. These side effects may include nausea, vomiting, drowsiness, anorexia, weight loss and constipation. A study of cancer patients by Smith et al.,25 however, indicated that these adverse effects occurred at statistically lower rates in IOT patients than in patients receiving comprehensive medical management, which consisted primarily of oral opioid analgesics. Interestingly, the authors also found that the patients receiving IOT also lived longer than did those in the comprehensive medical management group. The incidence of sexual side effects among patients receiving IOT is extremely high. Several studies26-31 have associated IOT with hypothalamic-pituitary-gonadal suppression. Roberts et al.3 found that 71% of male and 48% of female IOT patients in their study reported decreased libido, with 59% of the males reporting erectile difficulties and 47% of the females under the age of 50 reporting menstrual abnormality. While the provision of gonadal steroids has been recommended,26,31 it is unclear whether practitioners providing IOT are routinely doing so. Accordingly, the possibility that serious and demoralizing sexual side effects of IOT go untreated exists.

The development of tolerance to intrathecally-administered opioids in long-term therapy has been expressed as a concern in the literature.5,32 However, a number of studies have suggested that tolerance is typically not problematic to the extent that it precludes continued IOT. Winkelmuller and Winkelmuller33 found that tolerance developed in only 7 of 120 nonmalignant chronic pain patients receiving IOT who were followed for an average of 3.4 years, with only 3 of the patients requiring removal of their pump. Other investigators of intrathecal administration of opioids to both cancer and non-malignant chronic pain patients34,20,10,35,36,3,37 have similarly found that tolerance to IOT is generally developed very gradually following initial dose titration, and is not likely to significantly interfere with long-term treatment. The use of intrathecally-administered adjuvants such as bupivicaine, lidocaine, clonidine, ziconotide, and neostigmine have been reported to reduce problems associated with tolerance.19,38 Opioid dosage escalation, of course, is considered to be a more salient issue in IOT of nonmalignant chronic pain than in the treatment of pain due to cancer.

While IOT is generally thought to be superior to systemic administration of opioids from an abuse perspective, two case reports indicate that abuse potential still exists. Gock et al.39 presented a report of a back pain patient with a history of opiate abuse who fatally overdosed on morphine removed from his pump and injected intravenously. Burton and his colleagues40 presented a case study of an incarcerated patient receiving IOT for chronic back pain whose pump contained phencyclidine, methamphetamine and propoxyphene in addition to the morphine administered by his treating physician. Although rare, cases of substance abuse from opioids obtained from an implanted intrathecal pump as well as via an intrathecal pump have been documented.

Despite the rare cases of abuse associated with IOT cited above, intrathecal administration of opioid analgesics is becoming more widely accepted as an effective alternative to systemic opioids in the treatment of chronic pain. Nevertheless, IOT is invasive, prone to side effects and complications mentioned earlier in this article, costly (although potentially cost-efficient), and requires considerable technical support. As is true for any type of treatment for chronic pain, some patients are going to respond better to IOT than are others. Accordingly, selection of the chronic pain patients who are going to receive IOT has received considerable attention in the literature.

Psychological Assessment

Block’s41 landmark work on presurgical psychosocial screening prior to elective surgeries provides a highly systematic method for determining which patients are likely to benefit from surgical intervention versus those who are at risk for failure. Unfortunately, the number of surgeons who take advantage of Block’s selection system outside of teaching hospitals is questionable. A review of the literature fails to yield any investigations of the extent to which Block’s algorithm and decision-making process has been implemented. Despite empirical evidence of successes of IOT for chronic pain management, skepticism regarding the procedure continues to exist due to potential problems with complications, side effects and substance abuse. Unfortunately, an empirically-based decision-making process for IOT selection similar to Block’s41 process for screening elective surgery patients has yet to be developed. In fact, the literature on selection of appropriate candidates for IOT based upon psychological factors is sparse and quite limited in terms of empirical support. As recently as 2000, Doleys42 noted that none of the psychological variables associated with positive outcome of IOT have emerged through a process of scientific inquiry, with practitioners relying primarily upon conjecture and hypotheses based upon other types of treatment. Accordingly, the physicians who implant and manage intrathecal opioid pumps are as likely to be as challenged by issues of selection as by the mechanics of implantation and the management of complications and side effects.

Several investigators and reviewers43,42,44-49 of IOT in the treatment of chronic pain have emphasized the importance of psychological assessment prior to its initiation. Despite the growing number of intrathecal opioid pumps being which are being implanted, there exists a paucity of literature regarding what should be involved in providing such evaluations. A review of the literature indicates that the role of psychological evaluation in the implantation of spinal cord stimulators has been investigated considerably more thoroughly than has the role of psychological evaluations in IOT. Reviews42,50 have discussed the role of psychological assessment in “implantable” and “interventional” therapies, lumping IOT and spinal cord stimulators together to a certain extent. Doleys and Dinnoff50 recently noted that little attention has been paid to the exploration of the role of psychological factors impacting outcome of IOT relative to the amount of research on psychological factors impacting the outcome of spinal cord stimulation. While certain psychological aspects of these types of treatments may be the same, it is important to note that the potential side effects, complications and the possibility of substance abuse from or through an intrathecal opioid pump would suggest that the purpose of psychological screening prior to IOT may differ from that prior to spinal cord stimulator implantation.

Certain psychological factors are considered absolute contraindications for IOT. These include severe cognitive deficits (e.g. due to dementias) and frank psychosis.47 However, other exclusionary criteria which have been cited have included severe characterological disturbance, unremitting somatization, relapsing chemical dependency, malingering or factitious elements, unresolved, difficult litigation,48 and major depression.51 IOT is clearly contraindicated in patients suffering from dementia and psychoses due to the high likelihood of self-harm. Additionally, few would argue that the identification of malingering or factitious elements should not serve as a contraindication to the provision of IOT. However, the other contraindications which have been listed are more questionable, as the extent to which they are present should be taken into consideration prior to excluding patients from intrathecal pump implantation.

For example, while several authors5,48,49 have recommended excluding characterologically-disturbed patients from IOT, a review of the literature indicates that the relationship between personality disorders and outcome of IOT has yet to be studied. In two reviews, however, Schatman52,53 concluded that the most “dramatic” characterologically disturbed chronic pain patients do not fare well in any type of pain management treatment. It has been well established that characterologically-disturbed patients are overrepresented in the chronic pain population, with studies54,55 suggesting that as many as 50% of chronic pain patients suffer from a diagnosable personality disorder. The diagnosis of personality disorders is unfortunately highly subjective, with the dynamic between the patient and the health care provider often the primary determinant of the provision of such a diagnosis and the assessment of its severity. Unfortunately, patients who irritate health care providers are more likely to be determined to be severely characterologically disturbed, and accordingly may be denied IOT. In terms of psychological evaluation, however, it is questionable whether personality-disordered chronic pain patients should necessarily be denied IOT until empirical findings supporting such exclusion are presented.

“Somatization” can range from mild to extreme, with extreme somatization constituting a diagnosable Somatization Disorder. According to the Diagnostic and Statistic Manual of Mental Disorders (4th ed.) (DSM-IV).56 A Somatization Disorder is not diagnosable unless the patient presents with symptoms without medical explanation. Clearly, IOT is contraindicated in patients whose pain is not due to an identifiable physical source.45,57 However, “somatization” is a term commonly used to describe a process whereby physical symptoms, including pain, are exacerbated in response to stress. Patients prone to somatization are thought to have inadequate cognitive coping skills,58 with cognitive coping skills training found to be effective in terms of reducing subjective pain experience.59-61

“Chemical dependency” and “drug addiction” are also somewhat subjective concepts which are considered to be contraindications for IOT. Several investigators48,38,49 have suggested using current problems with dependence or abuse as a contraindication for IOT, while others have recommended ruling out IOT for patients who present with what are vaguely described as “inappropriate drug habituation”48 or “addictive personalities.”5,57 A current review of the literature indicates that the relationship between drug dependence, abuse or addiction and outcome of IOT has yet to be studied. While few would recommend IOT for a chronic pain patient who presents with a persistent and ongoing illicit drug habit, one of the problems with excluding anyone thought to have a “problem” with substances is that the potential for pseudoaddiction among chronic pain patients is extremely high.62-65 Perhaps a systematic investigation of the impact of IOT on the drug-seeking behavior of under-medicated chronic pain sufferers would determine whether appropriately titrated doses of intrathecally-administered opioids would potentially result in the discontinuation of or at least a significant reduction in aberrant drug-related behavior among certain chronic pain patients. Portenoy and Savage32 noted that “Addiction to opioid medications caused by pain treatment is very rare in individuals with no history of addictive disorders”. They also noted, however, that while recovering addicts may be less likely to relapse with IOT than with systemic opioid administration, IOT provides these patients with access to quantities of highly concentrated opioids. The case studies of recreational abuse of drugs from,39 and through,40 intrathecal pumps which were mentioned earlier in this article certainly support Portenoy and Savage’s32 concern. However, the function of “drug-seeking behavior” among chronic pain sufferers being considered for IOT ought to be carefully assessed prior to excluding them from treatment.

There exists a considerable amount of disagreement in the literature regarding the extent to which ongoing litigation relating to personal injury is likely to impact the efficacy of IOT. Investigators and reviewers48,38 have listed compensation or ongoing litigation as a psychological exclusion criterion for chronic IOT for nonmalignant pain, with another author57 recommending that caution be exercised in cases involving ongoing litigation or an insurance claim. These recommendations are supported, to some extent, by a recent study14 which determined that IOT patients who were not receiving Workers’ Compensation demonstrated statistically significant improvements in level of perceived disability at 12-month follow-up while those on Workers’ Compensation did not experience significant gains. This finding is contradicted, however, by several studies27,66,49 which failed to identify significant differences in pain levels or improvement in activity levels between Workers’ Compensation and non-Workers’ Compensation IOT patients. A number of investigators3,34,36 have determined that IOT does not result in significant changes in work status. These mixed results should not be used to exclude patients who are involved in litigation or are receiving wage loss indemnity from IOT, as these patients do not necessarily represent a homogeneous group. Specific research on the impact of secondary gain on IOT outcome would help determine which patients who are receiving or seeking some type of financial compensation are at risk for suboptimal results.

The issue of depression as an exclusionary criterion for IOT is an extremely complex one. Several investigators and reviewers44,67,51 have listed depression or “mood disorders” as a contraindication for IOT, while others5,49,57 have vaguely listed a “psychiatric disorder” or “psychiatric disturbance” as a precluding factor. The literature citing the inordinately high prevalence of depression among chronic pain patients is extensive, with estimates as high as 100% reported.68 While reported rates of depression among chronic pain patients have been found to range from 10% to 100%,69 most studies have reported depression in over 50% of their chronic pain patient sample.70 Certainly, methodological issues have impacted the considerable variance which investigators have reported. It should be noted that studies of depression and chronic pain have generally focused on major depressive disorders, and have accordingly failed to identify chronic pain sufferers who experience considerable depressive symptomatology yet do not quite reach diagnosable levels of depression based upon DSM-IV criteria. It is also important to recognize that the severity of chronic pain has been empirically related to the severity of depression.71-74 Given that chronic pain patients who present as candidates for IOT are likely to be experiencing severe subjective physical discomfort, it would accordingly be surprising not to see them present as clinically depressed. While the potential for self-harm associated with IOT should preclude patients who are actively suicidal,38 depression without suicidality should not be considered a contraindication for IOT until empirical evidence indicates that depressed patients do not respond well to this type of pain management approach.

“Given that chronic pain patients who present as candidates for IOT are likely to be experiencing severe subjective physical discomfort, it would accordingly be surprising not to see them present as clinically depressed.”

Given the potential complications and side-effects associated with IOT which were discussed earlier in this article, what is more important than the presence of depression in the psychological selection of candidates for intrathecal opioid administration are the chronic pain patient’s coping skills and expectations regarding treatment outcome. Several investigations and reviews43,42,75 have recommended that certain psychodiagnostic tests be utilized in the screening of potential IOT candidates. Unfortunately, empirical evidence regarding the ability of these tests to predict success versus failure of IOT is lacking. Only a study by Doleys and Brown75 utilized psychological testing in an effort to predict outcome of IOT, with results suggesting that the Minnesota Multiphasic Personality Inventory-II (MMPI) may not be a particularly good predictor, with the “presence of a relatively normal MMPI profile” not found to be predictive of a more positive outcome as compared to a more abnormal profile. Generalizing the results of investigations which used psychological testing to determine the outcome of spinal cord stimulators to IOT is not necessarily sound practice, as an empirical basis for doing so has not yet been established. It is important to note, however, that the MMPI is still likely to be useful in screening potential IOT recipients, as its strengths are the identification of malingering through the validity scales and the identification of psychoses which relatively well-defended patients may be able to obscure in a clinical interview.

While there may be some benefit to administering an MMPI, the key to the psychological evaluation of appropriateness for IOT is thorough assessment through a clinical interview. Direct questioning of a patient regarding his or her expectations of the likely benefits as well as the potential difficulties associated with IOT should provide the treatment team with crucial information regarding the likelihood that the patient will consider IOT to be successful. Since the largest multicenter study of the efficacy of IOT for cancer and noncancer pain conducted to date36 found mean relief of 61% after an average of 14.6 months of treatment, IOT is contraindicated for patients who are adamant regarding their expectations of complete or even near-complete pain relief. Significant others should be included in the screening interview, as their expectations regarding outcome are important as well, particularly given the finding that pain-relevant spousal support can serve as a buffer against depression among chronic pain patients.76

Another important aspect of the clinical interview prior to IOT is the assessment of a patient’s likely emotional and behavioral responses to the potential complications and side-effects of IOT. The psychologist should be forthright in sharing the distinct possibility that there exists a considerable likelihood that complications will require additional surgery, with an effort made to determine whether this will temper the patient’s enthusiasm and motivation for IOT. As IOT is considered a “treatment of last hope,” the psychologist should determine whether stressful and disappointing complications will likely lead to the development of hopelessness, which has been associated with medical non-compliance.77

Sexual dysfunction, as mentioned above, is an extremely common side effect of IOT, although appears to be no more pronounced than among chronic pain patients taking high dosages of oral opioids.78-79 Nevertheless, as sexual side effects of various other types of medications80-82 have been associated with noncompliance, the evaluating psychologist should assess the potential emotional and behavioral impact of opioid-related sexual dysfunction for each patient, as well as for his or her significant other.

Monsalve et al.83 developed an algorithm for the selection of appropriate candidates for the implantation of neuromodulation systems, which they applied to IOT as well as spinal cord stimulators. The authors listed a formidable group of exclusionary criteria, which included severe cognitive deficits, severe addictive behaviors, psychotic or schizophrenic disorders, high level of anxiety or depression, a clear profile of somatization or conversion disorder, severe sleep disturbances, personality disorders, passive coping styles, “catastrophism”, and “assimilative coping styles.”83 To support these exclusionary criteria, the authors cited numerous articles and book chapters. Interestingly, however, none of their supporting literature was empirically-based, and none pertained specifically to IOT. Accordingly, the validity of their algorithm cannot be considered particularly strong, and its exclusion of such a wide range of chronic pain patients from IOT should be considered arbitrary and irresponsible. Monsalve and colleagues83 presented an application of their algorithm, comparing patients who did not present with the aforementioned exclusionary criteria to patients who were not assessed through their algorithm. Not surprisingly, they did not distinguish between IOT and spinal cord stimulator recipients in their data analysis, making the relatively small difference in outcome (which was measured solely through Visual Analogue Scale score change) even less meaningful.


In summary, the exclusion of potential IOT candidates based upon psychological factors needs to be considered carefully, particularly until there exists empirical data upon which such exclusion is based. While the body of empirical data supporting the exclusion of patients from the implantation of spinal cord stimulators is limited, empirical support for the exclusion of chronic pain patients from IOT is essentially non-existent. With the exception of dementias, psychoses, ongoing addiction and suicidal depression which could result in a heightened risk of self-harm along with the relatively rare malingering and factitious disorders, very few physically appropriate patients should be excluded from IOT based upon a single psychological factor. Until more research on psychological factors and IOT outcomes is presented, the psychological evaluation of IOT candidates should be considered more of an art than a science. Psychologists who are experienced in evaluating chronic pain patients as a part of an interdisciplinary team should consider each candidate individually and phenomenologically, taking the patient’s expectations of the impact of IOT on the overall quality of his or her life into the highest consideration when making recommendations to the treatment team. While psychological testing can provide information regarding the potential benefits of adjunctive or preparatory treatments such as psychopharmacotherapy, individual or family counseling, or biofeedback/relaxation training, the lack of empirical evidence supporting prediction of IOT results through psychological testing suggests that testing data should be applied cautiously.

IOT represents an extremely appealing treatment option for many patients whose chronic pain has severely detracted from their overall quality of life. While physicians should avoid the pitfall of ignoring the psychological factors which may put recipients of intrathecal opioid delivery systems at considerable risk, psychologists working as members of interdisciplinary pain management teams need to recognize that their evaluation of these patients should serve only to enhance outcomes. Excluding patients from IOT based upon a “science” which has not yet been fully established puts chronic pain sufferers at risk of losing access to a treatment which may represent their only chance for comfort and a reasonable quality of life. Psychological testing, while of some benefit, has not yet been determined to differentiate between successful and unsuccessful intrathecal opioid therapy recipients. Accordingly, pain management specialists need to exercise caution in excluding patients whose physical conditions are being affected by psychological factors, with the exception of a small group of gross contraindications.

Last updated on: July 18, 2013
close X