SEARCH

SEARCH BY CITATION

Keywords:

  • cognitive dysfunction;
  • delirium;
  • cancer;
  • pain;
  • opioid metabolites;
  • hydration;
  • renal impairment;
  • aging;
  • opioid switch;
  • opioid toxicity

Abstract

  1. Top of page
  2. Abstract
  3. GENERAL CONTEXT OF OPIOID-RELATED COGNITIVE DYSFUNCTION IN PATIENTS WITH CANCER
  4. PREVALENCE AND PATTERN OF COGNITIVE EFFECTS ASSOCIATED WITH OPIOID USE
  5. DELIRIUM AND COGNITIVE DYSFUNCTION: THE ROLE OF OPIOIDS
  6. PATHOPHYSIOLOGY OF OPIOID-ASSOCIATED COGNITIVE DYSFUNCTION
  7. IMPACT OF OPIOID-ASSOCIATED COGNITIVE EFFECTS ON PATIENT, FAMILY, AND STAFF
  8. ASSESSMENT OF COGNITIVE DYSFUNCTION IN OPIOID-TREATED PATIENTS
  9. MANAGEMENT APPROACH TO OPIOID-ASSOCIATED COGNITIVE IMPAIRMENT
  10. CONCLUSIONS
  11. REFERENCES

BACKGROUND

Opioids have an essential role in the management of pain in cancer patients, particularly those with advanced disease. Cognitive dysfunction is a recognized complication of opioid use. However, misconceptions and controversy surround the nature and prevalence of its occurrence. A projected increase in the aging cancer population highlights the need for a better understanding of this phenomenon.

METHODS

A critical appraisal of the literature evidence in relation to the pattern, pathophysiology, assessment, impact, and management of cognitive dysfunction due to opioid use in cancer pain management is given.

RESULTS

Studies in cancer patients with less advanced disease reveal subtle evidence of cognitive impairment, largely related to initial dosing or dose increases. In advanced cancer, opioid-induced cognitive dysfunction usually occurs in the form of delirium, a multifactorial syndrome. The presence of both cognitive impairment and delirium frequently is misdiagnosed or missed. Potential risk factors include neuropathic and incidental pain, opioid tolerance, somatization of psychologic distress, and a history of drug or alcohol abuse. Elevation of opioid metabolites with renal impairment may contribute to cognitive dysfunction. Recognition of opioid-related cognitive dysfunction is improved by objective screening. Successful management requires either dose reduction or a change of opioid, in addition to addressing other reversible precipitants such as dehydration or volume depletion.

CONCLUSIONS

Opioid-related cognitive dysfunction tends to be subtle in the earlier stages of cancer, whereas delirium, a more florid form with behavioral disturbance is likely to be present in the advanced cancer population. In patients with advanced disease, an optimal management approach requires careful clinical assessment, identification of risk factors, objective monitoring of cognition, maintenance of adequate hydration, and either dose reduction or switching to a different opioid. Cancer 2002;94:1836–53. © 2002 American Cancer Society.

DOI 10.1002/cncr.10389


GENERAL CONTEXT OF OPIOID-RELATED COGNITIVE DYSFUNCTION IN PATIENTS WITH CANCER

  1. Top of page
  2. Abstract
  3. GENERAL CONTEXT OF OPIOID-RELATED COGNITIVE DYSFUNCTION IN PATIENTS WITH CANCER
  4. PREVALENCE AND PATTERN OF COGNITIVE EFFECTS ASSOCIATED WITH OPIOID USE
  5. DELIRIUM AND COGNITIVE DYSFUNCTION: THE ROLE OF OPIOIDS
  6. PATHOPHYSIOLOGY OF OPIOID-ASSOCIATED COGNITIVE DYSFUNCTION
  7. IMPACT OF OPIOID-ASSOCIATED COGNITIVE EFFECTS ON PATIENT, FAMILY, AND STAFF
  8. ASSESSMENT OF COGNITIVE DYSFUNCTION IN OPIOID-TREATED PATIENTS
  9. MANAGEMENT APPROACH TO OPIOID-ASSOCIATED COGNITIVE IMPAIRMENT
  10. CONCLUSIONS
  11. REFERENCES

Opioids play a major role in the pharmacotherapy of pain control in patients with cancer.1, 2 Cognitive impairment is a frequent finding in this population3 and occurs in most before death.4–6 The specific contribution of opioids to the cognitive impairment associated with advanced cancer is often difficult to evaluate owing to the frequent presence of multisystem impairment, and the concurrent administration of other psychotropic medications.7 Furthermore, in patients with cancer, cognitive impairment frequently is associated with other neuropsychologic complications such as somnolence, perceptual disturbance, and mood alteration. These can occur independently, or perhaps more often collectively, within the multifactorial syndrome of delirium.4, 8 This syndrome is generally attributable in varying degrees to the cancer disease process and its complications, and also to the administration of opioids and other psychotropic medications.4

In the U.S. approximately 60% of all cancers occur in the population aged 65 years and older, and the predicted increase in this age group will likely be accompanied by an increase in cancer incidence and cancer-related mortality.9–11 This trend warrants greater physician awareness and understanding of the nature of opioid-related cognitive dysfunction. Pain is a highly prevalent symptom in the elderly before death.12–14 There is evidence to suggest that opioids are underutilized in the elderly.15 Although awareness and vigilance on the part of patient and physician regarding the cognitive side effects of opioids is appropriate, undue fear could contribute to so-called opiophobia and opioid underutilization.

In view of the multiple confounding variables, it therefore is not surprising that there is a paucity of literature studies that selectively evaluate the cognitive effects of opioids in the cancer population. A pragmatic conceptual approach in this situation recognizes cognitive impairment as commonly constituting part of the global syndrome of delirium, along with other required criteria. Episodes of opioid-induced cognitive impairment without other delirium diagnostic criteria can in the absence of dementia represent subsyndromal delirium.16, 17 Cognitive functioning refers to the brain's acquisition, processing, storage, and retrieval of information. This is not synonymous with alertness, although a certain level of alertness is necessary to allow assessment of cognitive function, and the two therefore are often studied in combination.18 To selectively examine the cognitive effects of opioids and avoid the many confounding variables present in the advanced cancer population, studies have been conducted in healthy volunteers,19–25 patients with chronic nonmalignant pain,26–29 and in cancer patients with less advanced disease.30 Some studies of cognitive dysfunction in advanced cancer patients have attempted to specifically examine opioid effects, particularly in relation to cognitive or psychomotor deficits,18, 31–37 or in relation to delirium.38 Other studies have studied cognitive impairment or delirium from a broader perspective and have variably examined the occurrence rates, the multifactorial etiology, and management aspects of these phenomena.4, 5, 39, 40

This review explores the literature evidence documenting the prevalence and pattern of opioid-associated cognitive effects and then addresses the pathophysiology, impact, assessment, and specific management of opioid-associated cognitive impairment, largely within the context of delirium.

PREVALENCE AND PATTERN OF COGNITIVE EFFECTS ASSOCIATED WITH OPIOID USE

  1. Top of page
  2. Abstract
  3. GENERAL CONTEXT OF OPIOID-RELATED COGNITIVE DYSFUNCTION IN PATIENTS WITH CANCER
  4. PREVALENCE AND PATTERN OF COGNITIVE EFFECTS ASSOCIATED WITH OPIOID USE
  5. DELIRIUM AND COGNITIVE DYSFUNCTION: THE ROLE OF OPIOIDS
  6. PATHOPHYSIOLOGY OF OPIOID-ASSOCIATED COGNITIVE DYSFUNCTION
  7. IMPACT OF OPIOID-ASSOCIATED COGNITIVE EFFECTS ON PATIENT, FAMILY, AND STAFF
  8. ASSESSMENT OF COGNITIVE DYSFUNCTION IN OPIOID-TREATED PATIENTS
  9. MANAGEMENT APPROACH TO OPIOID-ASSOCIATED COGNITIVE IMPAIRMENT
  10. CONCLUSIONS
  11. REFERENCES

Folstein et al. reported cognitive impairment in cancer patients with a prevalence range of 14–29% in a consecutive admission series.39 In hospitalized cancer patients, who were on opioids and referred to a supportive care service, Leipzig et al. found that 77% (27 of 35) experienced “impaired mental status” at least once when observed prospectively for a minimum of 7 days.40 The impaired mental status was based on an impaired assessment on a brief modified portion of the Minimental State Examination.41 The frequency of cognitive impairment on objective testing in patients with cancer on admission to a palliative care unit has been reported in the range of 20–44%,3, 42–44 whereas 80–90% show impairment before death.4–6 The average duration of cognitive impairment before death in a prospective study of patients with advanced cancer has been estimated (the sample was not an inception cohort) to be 16 days.5 A prospective study of a similar population reported that delirium was present for a median of 6 days (range, 1–47) before death.4 Collectively, these data clearly attest to the high frequency of cognitive impairment in patients with advanced cancer. Although opioid use is likely to be high in this patient population, other etiologic factors for cognitive impairment, including the presence of dementia need to be considered. Prospective studies that examined the occurrence rates of cognitive dysfunction in cancer patients are discussed further in association with causal and management aspects and are summarized in “Management Approach to Opioid-Associated Cognitive Impairment”. Heterogeneity of patient sampling and diagnostic criteria hinder the conduct of a meta-analysis of the occurrence rates of cognitive dysfunction in cancer patients.

Studies of opioid-related cognitive dysfunction in human volunteers are summarized in Table 1. In a study of opioid-naïve, healthy volunteers, a computerized steady state infusion of morphine compared with a saline infusion resulted in increased reading time and impairment of later recall of previously read textual material.19 Other studies examining the psychomotor and cognitive effects of opioids in opioid-naïve, healthy volunteers have largely suggested morphine-associated decrements in speed of reaction but either minimal or no reduction in accuracy.20–22 A recent study suggested that psychomotor and cognitive impairment in the opioid-naïve, healthy volunteer population was greatest in relation to meperidine, less in relation to hydromorphone, and least of all in relation to morphine.23 Impairment of cognitive and psychomotor testing appears to be more pronounced when an opioid is administered parenterally,24, 25 as opposed to orally.20, 21 Furthermore, opioid-related impairment appears to be dose-related,22 but in general the magnitude of impairment is markedly less than that associated with lorazepam.20, 21 Again, caution must be exercised when extrapolating the findings of these studies to the cancer population, on the grounds that volunteers are healthy and in no pain.

Table 1. Studies of Cognitive Dysfunction or Psychomotor Impairment after Experimental Opioid Administration in Healthy Volunteers
ReferenceNo. of PatientsM/F ratioAge (range)DesignOpioid doses and administrationMain cognitive or psychomotor performance measuresPrincipal findings
  1. M/F: male/female; RDBPCT: randomized double blind placebo-controlled trial, XO: crossover feature; Mor: morphine; i.v.: intravenously; p.o.: orally; Dxt: dextropropoxyphene; HM: hydromorphone; Oxy: oxycodone; i.m.: intramuscularly; Bup: buprenorphine; CFFT: critical flicker fusion test; DSST: digit symbol substitution test; VAS: visual analog scale.

Kerr et al. (1991)1915All males21–37RPCTMor vs. saline. Mor i.v. to achieve low, medium, and high plasma levelsTime needed to encode and process serially presented verbal information. Delayed recall of information.Time needed to encode and process verbal information was greater with Mor. Delayed recall was significantly impaired with Mor.
Hanks et al. (1995)20128/430–47RDBPCT XO designMor 10m and 15 mg p.o., lorazepam 1 mg p.o., (positive control) and placebo p.o.Reaction times, number vigilance, memory scanning, word recall, word recognition, picture recognition, CFFT, and subjective measures of alertness, calmness, and contentment.Both Mor doses produced impairment in delayed word recall and picture recognition. Significant improvement in choice reaction times with Mor 15 mg. Lorazepam caused marked impairment in tests of attention and memory.
O'Neill et al. (2000)21104/625–40RDBPCT XO designMor 10 mg p.o., Dxt 100 mg p.o., lorazepam 1 mg p.o. (positive control) and placebo p.o.Reaction times, number vigilance, memory scanning, word recall, word recognition, picture recognition, CFFT, and subjective measures of alertness, calmness, and contentment.Neither opioid had any substantial effects on cognition compared with lorazepam.
Zacny et al. (1994)221210/221–33RDBPCT XO designMor given i.v. as 0, 2.5, 5, or 10 mg/70 kgDSST and reaction times. Subjective ratings of sedation, confusion etc. using VASImpaired DSST with Mor. Most effects of Mor were dose-related.
Walker and Zacny (1999)231610/621–32RDBPCT XO designGraded i.v. doses of HM, Mor, or MepDSST, logical reasoning test. Subjective ratings of sedation, confusion etc. using VASDose-related impairment of DSST with all three opioids. Mildest impairments were with Mor. Logical reasoning was not affected.
Saarialho-Kere et al. (1989)24910/620–26RDBPCT XO designOxy 0.13 mg/kg i.m., inactive placebo p.o. saline i.m., and diphenhydramine 100 mg p.o.DSST and reaction times. Divided attention test. Subjective ratings of “sedation,” “confusion,” etc. using VASTransient impairment of reaction times and attention testing with Oxy. There was a significant subjective increase in “mental slowing” and feeling “muzzy” with both Oxy and diphenhydramine.
MacDonald et al. (1989)2512All males19–38RDBPCT XO designBup 0.3 mg i.m. ketorolac 30 mg. i.m., diclofenac 50 mg i.m. and placebo i.m.Battery of eight tests, including choice reaction times, digit span, CFFT, computerized test of attention, and VAS for subjective central effectsBup caused significant impairment in 7 of 8 of the tests with effects peaking consistently 4 hours after dosing.

Studies examining opioid-related cognitive dysfunction in chronic nonmalignant pain are limited, and overall they provide a conflicting series of results (Table 2). Lorenz et al. reported an improvement in perceptual–cognitive status when six opioid-naïve chronic nonmalignant pain patients were started on oral morphine (30–150 mg/day) treatment.26 This improvement occurred in parallel with their pain relief. Moulin et al. conducted a randomized controlled trial of oral morphine in chronic nonmalignant pain patients who previously were exposed only to weak opioids.27 Doses of up to 120 mg of morphine per day resulted in no decrease in cognitive testing scores using the High Sensitivity Cognitive Screen,45 although this appears to have been administered on a relatively infrequent basis and therefore could have missed a short episode of cognitive impairment. A prospective study of 19 chronic nonmalignant pain patients before and after stabilization on long-acting opioids showed no decline in cognitive functioning, as evidenced by objective testing with a battery of tests.28 Another prospective study in 40 patients with chronic nonmalignant pain versus age-matched control volunteers found significant cognitive and psychomotor impairment occurring in association with opioid treatment, using a median oral morphine dose of 60 mg (range, 15–300 mg) or equianalgesic doses of other opioids.29 However, the authors acknowledge the potential impact of anxiety and depression in the patients with pain, to adversely impact neuropsychologic performance. The presence of pain has been suggested by some authors to moderate the potentially negative influence of opioids on alertness and cognition.21 The presence of pain as an antagonist to the sedative opioid effects is tentatively analogous to the effect of pain on the respiratory depressant effects of opioids, in which opioid-associated respiratory depression emerges when the pain stimulus is removed.46–48

Table 2. Studies of Opioid-Related Cognitive or Psychomotor Dysfunction in Chronic Nonmalignant Pain
ReferenceNo. of patientsM/F ratioAge (yrs)DesignOpioid doses and administrationMain performance measuresPrincipal findings
  1. M/F: males/females; RDBPCT: randomized placebo-controlled trial; XO: crossover feature; Mor: morphine; SR: slow release; p.o.: orally; MEDD: morphine equivalent daily dose as mean ± standard deviation; HVLT: hopkins verbal learning test; DSST: digit symbol substitution test; CRT: continuous reaction time; FTT: finger tapping test; PASAT: paced auditory serial addition task; CI: cognitive impairment.

Lorenz et al. (1997)266All females38–61Case seriesMor SR p.o. (median; 35 mg; range; 30–150 mg)Reaction time before and during experimental painImproved reaction time on Mor
Moulin et al. (1996)274619/2722–67RDBPCT XO designMor SR 60 mg p.o. b.i.d. vs. benztropine 1 mg p.o. b.i.d.High sensitivity cognitive screenNo difference between Mor and active placebo
Haythornthwaite et al. (1998)2819 cases 10 controls8/11 2/8Non-RCTMEDD p.o. at follow-up: 111.1 mg (±69.6) in cases and 19.0 mg (±18.5) in controlsHVLT, trailmaking tests A and B, digit span, DSST, verbal learning testNo CI with use of opioids. Stable opioid doses associated with better test scores
Sjogren et al. (2000)2940 cases16/2446–74CTMor p.o. median 60 (range, 15–300) mgCRT, FTT, PASAT in cases and 40 healthy volunteers (controls)Poorer scores in the pain group. Memory tests and pain scores had a positive correlation, ? pain had an arousal effect.

Prospective studies in cancer patients that specifically examined opioid-related cognitive dysfunction are summarized in Table 3. Clemons et al. reported a small study of cancer patients, 7 on stable morphine doses, 6 on no morphine, and 16 healthy, age-matched controls.18 Here, all cancer patients scored poorer than healthy controls on cognitive assessments, and those on morphine performed the poorest, suggesting that the opioid contribution to the poorer cognitive performance was superimposed on the effect of the cancer disease process. Vainio et al. compared cognitive and psychomotor functioning related to driving ability in 24 cancer patients on stable morphine doses and in 25 pain-free patients, who were on no opioids.30 Although the morphine group performed more poorly, there was no significant difference between the two groups, suggesting that the use of long-term stable doses of opioid in cancer patients is not hazardous for driving.

Table 3. Prospective Studies of Opioid-Related Cognitive or Psychomotor Dysfunction in Patients with Cancer
ReferenceStudy population sampleOpioid doses and administrationMain performance measuresPrincipal findings
  1. M/F: males/females; Mor: morphine; p.o.: orally; epi: epidurally; i.m.: intramuscularly; CRT: continuous reaction time; DSST: digit symbol substitution test; FTT: finger tapping test; PASAT: paced auditory serial addition task; CI: cognitive impairment.

Bruera et al. (1989)32Stable opioid dose (n = 20), opioid dose increase (n = 20), similar demographicsPatients without a change in dose over 7 days vs. those with ≥30% increaseDigit and visual memory. Arithmetic test. Finger tappingImpairment of all cognitive and psychomotor tests after a dose increase of ≥30%.
Sjogren and Banning (1989)35n = 14 cases; age; 44–71 yrs; n = 20 healthy volunteer controlsMor p.o.: 130–400 mg, Mor epidural, 30–240 mgCRT scores in cases vs. those in controlsNo difference between those on p.o. versus epidural opioids. Prolonged CRT in both opioid groups compared with controls
Banning and Sjogren (1990)33n = 34 cases; M/F, 18/16; age, 33–71 yrsStable Mor oral doses; range, 30–920 mgCRT in cases vs. 32 healthy controlsSignificant prolongation of CRT in opioid-treated group
Banning et al. (1992)34n = 16 on opioids; M/F; 8/8, n = 16 controls; M/F, 8/12Mor p.o.: median, 150 (range 30–400) mg. Controls not on opioidsCRT in cases vs. controlsSignificant prolongation of CRT in opioid-treated group
Sjogren et al. (1994)36n = 22 on epidural opioids, n = 31 with cancer, p.o. opioids, n = 20 opioid-naive patients, n = 44 controls (no opioids)Mor Epi: 79 (12–600) mg (n = 22) Mor p.o.: 180 (30–920) mg (n = 31) Mor i.m.; 13 (9–23) mg single injection (n = 20).CRT in opioid groups vs. scores in 44 healthy volunteers (controls)Qualitative differences in prolongation of CRT between opioid-naive patients and controls
Vainio et al. (1995)30n = 24 cases; n = 25 controls (no pain and not on opioids)Mor p.o.: mean 209 (range, 60–1100) mgReaction times, nonverbal intelligence, and attention.Patients on Mor performed poorer (not statistically significant
Clemons et al. (1996)18n = 7 on stable opioid doses n = 6 on no opioid n = 16 age-matched controlsMor p.o. dose range: 50–200 mgMultiple tests including reasoning, reaction time, language, and memoryBoth cancer groups scored similarly. Authors concluded that CI was associated with advanced cancer and the opioid effects were superimposed.
Wood et al. (1998)31n = 18 patients admitted to hospice; M/F, 11/7; mean age, 59 (range, 29–76) yrsMor dose range: 40–600 mg (n = 9), and 10–120 mg (n = 9)Adult reading test, delayed recall, DSST, and the trailmaking testWeak correlation between plasma Mor concentration and scores on digits forward test and DSST. Unable to distinguish between disease and opioid effects
Sjogren et al (2000)37n = 130 cancer patients; age, 40–76, allocated in a cross-sectional design to 5 groups based on physical functioning, presence or absence of pain, and use or nonuse of opioidsMedian (range) doses: group with poor physical function and pain present, 120 (25–420) mg/24 hrs; group with poor physical function and no pain, 40 (20–180) mg/24 hrsCRT, FTT, PASATLong-term opioid use per se did not affect test performance. Authors concluded that pain and reduced physical functional status in association with opioid treatment may contribute to poorer performance on neuropsychologic testing.

Banning and Sjogren were involved in a series of studies that compared auditory continuous reaction times (CRTs) in cancer patients on opioids with those of non-opioid-treated cancer patients.33–36 One study concluded that CRT was greater in opioid-treated cancer patients, and there was a weakly positive correlation between analgesic dose and sedation visual analog with CRT.33 Another study found prolongation of CRT and higher sedation scores in a group of patients receiving opioids, either alone or in combination with other peripherally acting analgesics, versus those receiving peripherally acting analgesics alone.34 Both groups were matched for age and performance status in this study. An earlier study had demonstrated no difference in CRT between oral and epidural administration.35 Another study demonstrated qualitative differences in CRT between opioid-naïve patients treated with a single intramuscular dose of morphine and cancer patients on chronic oral or epidural morphine.36 However, there was no correlation between plasma morphine concentration and CRT values in any of the three groups, which is inconsistent with earlier findings. This highlights the need for further studies to examine the relation of cognitive deficits to route of opioid administration and also to levels of parent opioid or opioid metabolites.49 The most recently reported study of cancer patients by Sjogren et al. involved an elaborate cross-sectional design that aimed to assess the specific association of orally administered opioid, pain, and performance status with scores in a battery of neuropsychologic tests that assessed attention, psychomotor performance, and memory.37 Their findings suggested that although the combination of orally administered opioid, pain, and performance status impairment was associated with poorer scores in some aspects of neuropsychologic testing, long-term stable oral opioid dosing was not per se associated with any neuropsychologic test impairment. Impaired performance status alone was associated with poorer performance in attention testing, whereas pain alone was associated with poorer performance in memory testing.

In the clinical practice of cancer pain treatment, the goal of care involves achieving a balance between good pain control and minimal side effects, cognitive or noncognitive. The question therefore arises regarding the choice of opioid on initiation of treatment. The World Health Organization (WHO) ladder provides a guide to physicians in choosing an opioid: step 1 recommends a nonopioid and/or adjuvant analgesic for mild pain, step 2 recommends the addition of a weaker opioid such as codeine for mild to moderate pain, step 3 recommends the use of stronger opioids (such as morphine, hydromorphone, oxycodone, fentanyl, and methadone) in addition to nonopioid treatments and adjuvant analgesics.1 The weaker opioids such as codeine are commonly available in mixed drug formulations, for example, codeine and acetaminophen. This feature imposes a ceiling dose effect for analgesia. It is likely that the use of adjuvant analgesics such as nonsteroidal antiinflammatory drugs and steroids may have an opioid sparing effect,50 albeit that they too have a significant side effect profile, including cognitive impairment.51 Regarding the choice among the stronger opioids, there is little evidence to suggest that one opioid is superior to any other regarding cognitive side effects with the exception of meperidine and perhaps methadone. Use of meperidine is discouraged in the treatment of cancer pain, largely because of its neuroexcitatory effects.52 Compared with morphine in postoperative analgesia, meperidine is associated with greater cognitive impairment as measured by trailmaking tests.53 Methadone is largely reserved as a second-line stronger opioid, mainly because of its long and variable elimination half-life, and the consequent potential risk of significant cumulative toxicity.54 Although clinical experience suggests that cognitive impairment is less likely to occur with methadone, this has not been demonstrated in any prospective study to date.

A further question arises regarding the optimal opioid dose increments in the event of poor pain control. Dose increments in clinical practice are usually in the range of 30–50%.52 In the case of methadone, clinical experience suggests that in view of its longer half-life, increments might need to be more conservative and in the range of 20–25%. Bruera et al. compared a battery of 4 bedside cognitive tests, before and 45 minutes after their morning dose, in 2 groups of cancer patients on long-term opioid therapy.32 One group was exposed to stable opioid dosing over a week, and the other was exposed to a 30% or greater increase in opioid dose less than 3 days before study admission. The study findings suggested that cognitive impairment occurred immediately after a dose increase, and the authors concluded that tolerance develops to the cognitive effects based on the absence of cognitive impairment in the group receiving stable dosing.

A comparison of the studies that selectively examined the cognitive effects of opioids suggests several limitations relating to study methods but also enables us to make some simple conclusions. The sensitivity and sophistication of neuropsychologic testing is variable, sometimes simple and crude40 and at other times detailed and sophisticated.18 Variability in opioid doses, choice of opioid, dosing intervals, administrative routes, and timing of neuropsychologic testing in relation to opioid administration also must be recognized. Some studies have sought to examine the cognitive effects of opioids in more stable populations, either in chronic nonmalignant pain27–29 or in patients who are at a relatively early stage in the cancer disease trajectory.30 The findings of these studies are useful but clearly limited in their extrapolation to the advanced cancer population, in which cognitive impairment is more prevalent, vulnerability is higher, and there is an associated complex web of interaction between multiple etiologic factors.4

Overall, these studies suggest that sedation and cognitive impairment can occur in varying degrees with the initiation of opioid therapy and with dose increases of at least 30%. However, tolerance to these effects generally appears to develop in both of these situations. Chronic stable opioid dosing can at least be associated with minor cognitive effects, whose impact on patient functioning and distress is perhaps not yet fully determined. However, the best available evidence would suggest that chronic stable opioid dosing in cancer patients in itself should not necessarily require discontinuation of driving or work activity, if the patient is otherwise able to engage in these activities.30, 55 On the basis of the previously cited studies in healthy volunteers, the association of cognitive impairment appears to be greater when opioids are administered parenterally rather than orally, although the few studies examining the influence of administrative route in cancer patients failed to show this pattern. The confounding variables of the cancer disease process, its progression, the use of other psychotropic medication, patient anxiety, and mood disturbance must be borne in mind when evaluating the specific opioid contribution to cognitive deficits, especially in small study samples.

DELIRIUM AND COGNITIVE DYSFUNCTION: THE ROLE OF OPIOIDS

  1. Top of page
  2. Abstract
  3. GENERAL CONTEXT OF OPIOID-RELATED COGNITIVE DYSFUNCTION IN PATIENTS WITH CANCER
  4. PREVALENCE AND PATTERN OF COGNITIVE EFFECTS ASSOCIATED WITH OPIOID USE
  5. DELIRIUM AND COGNITIVE DYSFUNCTION: THE ROLE OF OPIOIDS
  6. PATHOPHYSIOLOGY OF OPIOID-ASSOCIATED COGNITIVE DYSFUNCTION
  7. IMPACT OF OPIOID-ASSOCIATED COGNITIVE EFFECTS ON PATIENT, FAMILY, AND STAFF
  8. ASSESSMENT OF COGNITIVE DYSFUNCTION IN OPIOID-TREATED PATIENTS
  9. MANAGEMENT APPROACH TO OPIOID-ASSOCIATED COGNITIVE IMPAIRMENT
  10. CONCLUSIONS
  11. REFERENCES

The multisystem impairment that accompanies progression of the cancer disease process is associated with an increase in vulnerability toward cognitive impairment, which is predominantly manifested as delirium.4, 56 Clearly defined criteria and clinical features are outlined in the delirium section of the fourth edition of the Diagnostic and Statistical Manual (DSM-IV).16 The main criteria for delirium are presented in Table 4. Delirium is characterized by a disturbance of consciousness with altered awareness, impaired attention, and a change in cognition or perceptual disturbance. The cognitive change in delirium can include memory impairment, disorientation, and language dysfunction with the caveat that this change is not better explained by a dementia. Perceptual disturbance includes misperceptions or illusory phenomena, and hallucinations, which are most commonly visual but also can include tactile, auditory, and olfactory types. Other clinical features of delirium include altered psychomotor activity, emotional lability, and sleep–wake cycle disturbance. Delirium typically has an acute onset over hours or days. The etiologic categories recognized in criterion D of the DSM-IV delirium diagnostic criteria include substance intoxication, substance withdrawal, a general medical condition, multiple etiologies, and indeterminate etiology.

Table 4. DSM-IV Diagnostic Criteria for Delirium Due to Multiple Etiologies
  1. Reprinted with permission from the Diagnostic and Statistical Manual of Mental Disorders, 4th ed. ©1994 American Psychiatric Association.

  • A
    Disturbance of consciousness (i.e., reduced clarity of awareness of the environment) with reduced ability to focus, sustain, or shift attention
  • B
    A change in cognition (such as memory deficit, disorientation, language disturbance) or the development of a perceptual disturbance that is not better accounted for by a preexisting, established, or evolving dementia
  • C
    The disturbance develops over a short period of time (usually hours to days) and tends to fluctuate during the day
  • D
    There is evidence from the history, physical examination, or laboratory findings that the delirium has more than one etiology (e.g., more than one etiologic general medical condition, a general medical condition plus substance intoxication or medication side effect)

The multifactorial nature of delirium was demonstrated in a recent study, which identified a median of 3 (range, 1–6) precipitating factors per episode of delirium in 71 cancer patients admitted to a palliative care unit.4 The median of incidental delirium was 3.5 (range, 1–22) days in the case of reversed episodes and 6 (range, 1–17) days in the case of nonreversed episodes. Analysis of precipitating factors revealed that opioids and dehydration were most commonly associated with delirium reversal. Although delirium reversed in 49% of episodes in this study, some patients had minor cognitive deficits or mild perceptual disturbance before or after an episode of delirium in the form of subsyndromal delirium. Although subsyndromal delirium has been studied in elderly general medical patients,57 it has not been the specific objective of studies in cancer patients to date.

Different subtypes of delirium have been described.58 These are classified as hypoactive, hyperactive, and mixed based on alteration in the level of psychomotor activity.59 Some authors have suggested that different etiologic factors might give rise to different subtypes.60 Furthermore, there is a suggestion that the outcome of delirium is more favorable in the “agitated” as opposed to the “somnolent” subtype.61 In the case of delirium associated with initiating opioid therapy, clinical experience suggests the hypoactive subtype is probably more common, in keeping with the early sedative effects of opioids. However, in the case of toxicity associated with chronic opioid use in cancer patients, the literature reports suggest a predominance of the hyperactive subtype or agitated delirium.62–65 It is certainly possible that opioid-induced hypoactive delirium is underdiagnosed, especially when regular cognitive screening is not conducted.66 Opioid-induced neurotoxicity has received increasing recognition in the last decade.17, 62, 65, 67–69 Cognitive impairment within the context of delirium is considered as one of these neurotoxic effects, which also include myoclonus, hyperalgesia, allodynia, perceptual disturbance, and seizures.

PATHOPHYSIOLOGY OF OPIOID-ASSOCIATED COGNITIVE DYSFUNCTION

  1. Top of page
  2. Abstract
  3. GENERAL CONTEXT OF OPIOID-RELATED COGNITIVE DYSFUNCTION IN PATIENTS WITH CANCER
  4. PREVALENCE AND PATTERN OF COGNITIVE EFFECTS ASSOCIATED WITH OPIOID USE
  5. DELIRIUM AND COGNITIVE DYSFUNCTION: THE ROLE OF OPIOIDS
  6. PATHOPHYSIOLOGY OF OPIOID-ASSOCIATED COGNITIVE DYSFUNCTION
  7. IMPACT OF OPIOID-ASSOCIATED COGNITIVE EFFECTS ON PATIENT, FAMILY, AND STAFF
  8. ASSESSMENT OF COGNITIVE DYSFUNCTION IN OPIOID-TREATED PATIENTS
  9. MANAGEMENT APPROACH TO OPIOID-ASSOCIATED COGNITIVE IMPAIRMENT
  10. CONCLUSIONS
  11. REFERENCES

The pathophysiology of opioid-induced delirium is complex, and largely still putative, not least because of the multifactorial input into delirium etiology. Nonetheless, developments in molecular biology and receptor pharmacology are providing new insights. Decreased serum anticholinergic activity has been demonstrated in general medical patients with delirium.70–72 However, a recent study of 22 elderly patients with a febrile illness showed an elevation in serum anticholinergic activity associated with the illness but not necessarily with the presence of delirium or medication changes.73 The anticholinergic activity associated with meperidine and its active metabolite normeperidine has been suggested as contributing to delirium and cognitive impairment associated with this opioid.74, 75

The specific role of opioid metabolites in the pathophysiology of opioid-induced neurotoxicity is unclear. Morphine is metabolized mainly to morphine-6-glucuronide (M-6-G), morphine-3-glucuronide (M-3-G), and to a lesser extent others, including normorphine and its glucuronides.76–79 M-6-G binds to opioid receptors and is a potent analgesic whereas M-3-G has poor affinity80 and therefore has no analgesic potency.81 The equianalgesic dose ratio of oral to parenteral morphine in single dose studies is 6 to 1, but in chronic oral dosing, this ratio changes to between 2 and 3 to 1, possibly related to enterohepatic circulation, which could facilitate an accumulation of M-6-G.82 Animal studies have demonstrated the induction of a neuroexcitatory state with M-3-G,83 normorphine-3-glucuronide,84 and also with the hydromorphone metabolite, hydromorphone-3-glucuronide.85

Opioid-induced neurotoxicity including cognitive dysfunction has been reported in the case of methadone,86 which has no known active metabolites, and in the case of an acute overdose of fentanyl,87 in which its temporal occurrence was likely to be inconsistent with a metabolite role in causation. Studies examining the correlation between the level of parent opioid or its metabolites and the level of cognitive deficit during opioid treatment have revealed some conflicting results. A study of cancer patients found no correlation between the presence of cognitive impairment and the M-6-G/morphine ratio.88 Vainio found a positive correlation between plasma concentrations of morphine and its glucuronides with impairment in neuropsychologic testing of attentive capacity, and of concentration and structuring ability.30 In a study of 18 cancer patients receiving morphine in a hospice setting, clinically inapparent cognitive impairment was revealed in a battery of cognitive tests.31 However, there was no correlation between the test results and morphine dose, plasma morphine levels, or M-3-G or M-6-G levels, although a weak correlation was noted between plasma morphine concentration and results of both digit forward and digit symbol substitution testing. In another study by the same authors, 19 of 36 hospice patients developed symptoms attributable to morphine treatment, nausea and vomiting in 10 and cognitive impairment in 9.89 All 19 patients with opioid-related side effects had significantly elevated creatinine levels, and compared with 17 patients without side effects, plasma M-3-G, M-6-G, and dose-corrected plasma M-3-G and M-6-G were significantly higher. The authors concluded that impaired renal function per se or elevated levels of morphine glucuronides due to renal impairment, or a combination of both, may have contributed to the opioid side effects.

Age is recognized as a risk factor for the development of delirium,90 and opioid use has been shown to be an independent risk factor for delirium in the elderly.91 Studies have demonstrated age-related elevation of morphine metabolites92 and reduction in volume of distribution of morphine.93 Aging is associated with a decline in renal function,94 and both polypharmacy and drug utilization in general is remarkably high in this segment of the population.95, 96 In general, it is likely that both altered pharmacokinetic and pharmacodynamic factors contribute to the risk of opioid-related cognitive dysfunction in this patient group.97 Several studies have documented the use of lower doses of opioids in the elderly.15, 98–100 Two of these studies were conducted specifically in palliative care settings, and their findings are unlikely to reflect unfounded physician fears or opiophobia but rather reflect the standard practice2 of titrating opioid dose to achieve adequate analgesia, a process limited by encountering toxicity.

The impact of renal impairment on opioid metabolite levels has been the object of several studies. These studies have consistently demonstrated elevated metabolite levels88, 89, 92, 101–106 when measured in renal impairment. They also have suggested a frequent association between renal impairment and opioid toxicity.65, 102, 106 The presence of dehydration as a contributory factor for opioid-associated cognitive impairment was suggested in a recent prospective study of delirium.4

IMPACT OF OPIOID-ASSOCIATED COGNITIVE EFFECTS ON PATIENT, FAMILY, AND STAFF

  1. Top of page
  2. Abstract
  3. GENERAL CONTEXT OF OPIOID-RELATED COGNITIVE DYSFUNCTION IN PATIENTS WITH CANCER
  4. PREVALENCE AND PATTERN OF COGNITIVE EFFECTS ASSOCIATED WITH OPIOID USE
  5. DELIRIUM AND COGNITIVE DYSFUNCTION: THE ROLE OF OPIOIDS
  6. PATHOPHYSIOLOGY OF OPIOID-ASSOCIATED COGNITIVE DYSFUNCTION
  7. IMPACT OF OPIOID-ASSOCIATED COGNITIVE EFFECTS ON PATIENT, FAMILY, AND STAFF
  8. ASSESSMENT OF COGNITIVE DYSFUNCTION IN OPIOID-TREATED PATIENTS
  9. MANAGEMENT APPROACH TO OPIOID-ASSOCIATED COGNITIVE IMPAIRMENT
  10. CONCLUSIONS
  11. REFERENCES

For the patient, cognitive impairment and the fluctuating symptomatology of delirium, including delusional and hallucinatory activity, often in association with psychomotor agitation, fear, and anxiety, impose a potentially major burden of distress.107 Cognitive impairment in the cancer patient, whether caused by opioids alone or in combination with other factors, imposes a major obstacle in symptom assessment, owing to the loss of the “gold standard” of reliable subjective patient input. This creates a major challenge for medical and nursing staff, particularly if there is associated agitation and therefore the risk of misinterpreting this as increased pain.108, 109 The patient's agitation and the inability of staff to communicate with the patient can lead to a reflex response on the part of some physicians, who might resort to a prompt increase in opioid dose, which, in turn, could further aggravate the situation, especially if the opioid is already contributing to the delirium.63 Cognitive impairment also introduces a major impediment to the process of patient counseling, often at a critical juncture in the cancer illness trajectory.110 Inability of the cognitively impaired patient to give informed consent often hinders participation in research studies. Finally, perhaps most compelling for the family or loved ones of patients with advanced disease, there is a reduced level of communication with the patient, often at a time when it is most precious.

ASSESSMENT OF COGNITIVE DYSFUNCTION IN OPIOID-TREATED PATIENTS

  1. Top of page
  2. Abstract
  3. GENERAL CONTEXT OF OPIOID-RELATED COGNITIVE DYSFUNCTION IN PATIENTS WITH CANCER
  4. PREVALENCE AND PATTERN OF COGNITIVE EFFECTS ASSOCIATED WITH OPIOID USE
  5. DELIRIUM AND COGNITIVE DYSFUNCTION: THE ROLE OF OPIOIDS
  6. PATHOPHYSIOLOGY OF OPIOID-ASSOCIATED COGNITIVE DYSFUNCTION
  7. IMPACT OF OPIOID-ASSOCIATED COGNITIVE EFFECTS ON PATIENT, FAMILY, AND STAFF
  8. ASSESSMENT OF COGNITIVE DYSFUNCTION IN OPIOID-TREATED PATIENTS
  9. MANAGEMENT APPROACH TO OPIOID-ASSOCIATED COGNITIVE IMPAIRMENT
  10. CONCLUSIONS
  11. REFERENCES

Some authors have questioned the significance of minor opioid-related cognitive deficits in the context of the cancer patient's daily functioning, particularly in the case of advanced disease.18 These “minor” deficits should not be dismissed too lightly. One concern is that minor cognitive deficits might herald the risk or onset of more profound deficits, and therefore this justifies the role of regular cognitive monitoring in these patients.3 Another concern is that previous studies have demonstrated that cognitive impairment is missed in 46% of elderly general medical inpatients,111 and similarly in 23–60% of cancer patients.5, 112 Misdiagnosis of delirium occurs in 5–40% of general medical referrals to psychiatric services for depression.113–115 One of the main reasons for missing the diagnosis of cognitive impairment or delirium is the failure to regularly conduct an objective cognitive assessment.66, 111 Cognitive impairment in the context of a mild hypoactive subtype of delirium is more likely to be missed.116 Cognitive screening of cancer patients with a briefly administered instrument such as the Minimental State Examination helps to detect cognitive deficits that otherwise might be missed.3 This instrument has been widely used, and its psychometric properties are well known. Furthermore, normal values have been calculated for age and educational level on the basis of a large population study.117 Explanation of the purpose of screening is often reassuring to those patients with inordinate fears of opioid-induced impairment of cognition.

Operationalizing DSM-IV diagnostic criteria, using a brief instrument such as the Confusion Assessment Method (CAM), in addition to the conduct of objective cognitive testing can make the diagnosis of delirium.118 Nursing staff can be trained to conduct the CAM. Other features of opioid-induced neurotoxicity, such as myoclonus, hyperalgesia, allodynia, and possibly tactile or visual hallucinations, should alert the physician to the possibility of opioid-associated cognitive dysfunction or delirium. The severity of delirium can be monitored using an instrument such as the Memorial Delirium Assessment Scale (MDAS)119 or the Delirium Rating Scale (DRS).120, 121 The MDAS was designed to serially monitor delirium in the cancer population. It incorporates behavioral items that are assessed on an observational basis and cognitive items that are objectively assessed, although 20–30% of advanced cancer patients are physically unable to directly undergo objective testing and therefore require prorating of scores.122 The cognitive, neurobehavioral, and other domains that are identified in the MDAS and DRS, along with their clinical manifestations, are summarized in Table 5. Instruments that are entirely observational also have been studied in the palliative care setting.4, 42, 123 Although these instruments are very applicable toward the end of life, their psychometric properties in general are likely to be inferior to those of the MDAS and DRS, which require patient participation in formal assessment of cognition.

Table 5. Domains Potentially Affected in Delirium and the Associated Clinical Manifestations
DomainClinical manifestation
Awareness and interaction118Reduced awareness of immediate environment and ability to interact
Cognitive capacity118, 119Deficits in attention, registration, orientation, and short-term memory
Thoughts and speech118Disorganized thoughts and incoherent speech
Perception118, 119Hallucinations, illusions, and misperceptions
Psychomotor activity118, 119Hypoactive, hyperactive, or mixed features of both
Delusions118, 119Paranoia sometimes occurs, but generally delusions are loosely held
Sleep pattern118, 119Reversal of the normal sleep–wake cycle
Emotional118Lability and disinhibition
Temporal119Acute onset, fluctuation in symptom intensity over a 24-hour period

Although objective testing remains the method of choice in evaluating and monitoring cognitive function, cancer patients who complain of cognitive impairment, often in the absence of objective evidence, warrant assessment in relation to depression, anxiety, and fatigue.124 A history of gradual onset of cognitive impairment over months or years, as opposed to hours or days, suggests the presence of dementia. Furthermore, a multidimensional assessment is essential, particularly with a view to identifying any underlying poor prognostic factors for pain control, such as a neuropathic pain component, incidental pain, opioid tolerance, somatization of psychologic distress, or a history of chemical coping through drug or alcohol abuse.125 Identification of some of these factors at an early stage could direct the physician to use more effective pharmacologic and nonpharmacologic intervention and therefore help prevent the occurrence of cognitive dysfunction associated with escalating opioid doses. Strategies for prevention of initial or recurrent episodes are summarized in Table 6.

Table 6. Proposed Strategies to Prevent Opioid-Induced Cognitive Dysfunction and Delirium
• Educate family regarding recognition of potential neurotoxic features such as myoclonus106
• Objectively monitor cognition108, 111
• Adjust dose in elderly patients96, 97
• Discontinue or minimize use of all unnecessary medications, especially other psychotropics4
• Adjust dose for impaired renal function93–95
• Maintain adequate hydration and use diuretics cautiously4
• Perform a multidimensional assessment127, 128
• Identify poor prognosticators for pain control: neuropathic and incidental pain, somatization, tolerance, and addiction history125
• Optimal utilization of other therapies e.g., appropriate use of adjuvant and other therapies, maximize nonpharmacologic interventions67
• Adopt a proactive approach: early intervention with management strategies to possibly avert a full-blown delirium syndrome17

MANAGEMENT APPROACH TO OPIOID-ASSOCIATED COGNITIVE IMPAIRMENT

  1. Top of page
  2. Abstract
  3. GENERAL CONTEXT OF OPIOID-RELATED COGNITIVE DYSFUNCTION IN PATIENTS WITH CANCER
  4. PREVALENCE AND PATTERN OF COGNITIVE EFFECTS ASSOCIATED WITH OPIOID USE
  5. DELIRIUM AND COGNITIVE DYSFUNCTION: THE ROLE OF OPIOIDS
  6. PATHOPHYSIOLOGY OF OPIOID-ASSOCIATED COGNITIVE DYSFUNCTION
  7. IMPACT OF OPIOID-ASSOCIATED COGNITIVE EFFECTS ON PATIENT, FAMILY, AND STAFF
  8. ASSESSMENT OF COGNITIVE DYSFUNCTION IN OPIOID-TREATED PATIENTS
  9. MANAGEMENT APPROACH TO OPIOID-ASSOCIATED COGNITIVE IMPAIRMENT
  10. CONCLUSIONS
  11. REFERENCES

The mainstay approach to management of opioid-associated cognitive impairment requires an appreciation of the potential underlying multifactorial etiology and the conduct of a disciplined multidimensional assessment.126, 127 In addition to addressing the opioid contribution to cognitive impairment, the physician must also recognize and treat the effects of other factors such as other psychotropic medications, dehydration, electrolyte disorder, infection, metabolic disorder, and hypoxia. The duration of delirium is dependent on the nature of the etiologic factors and also the presence or absence of an underlying dementia. Interventions aimed at an easily correctable factor can result in a relatively prompt resolution of delirium, in the absence of an underlying dementia.16 The presence of delirium in an advanced cancer patient often signals an ominous prognosis, yet reversal can be achieved in many situations. The level of intensity or aggressiveness applied to the various corrective measures for delirium and the consequent burden imposed on the patient is dictated by the goals of care.128 In this process, due recognition should be given to the inevitable presence of cognitive impairment or delirium in most patients in the last hours or days of life.4–6

Although any intervention in end-of-life care must be evaluated in terms of the risk benefit ratio, the management strategies directed specifically at opioid-induced cognitive dysfunction usually are associated with low risk and burden. Nonetheless controversy exists regarding the relative merits of opioid rotation vis-à-vis dose reduction in this situation.68, 129 The rationale of opioid switching or opioid rotation in the case of opioid-induced neurotoxicity is that a more favorable balance between analgesia and side effects is achieved on the newly substituted opioid, often at a lower dose than that predicted by the standard equianalgesic tables.2, 17, 130 It is speculated that in the context of opioid-induced delirium, switching to a different opioid allows achievement of analgesia on the newly substituted opioid, whereas the potentially toxic metabolites of the prior opioid are eliminated.17, 130 Although one might intuitively anticipate some amelioration of opioid-induced cognitive dysfunction when the opioid dose is reduced, and brief survey reports suggest so,131 this has not been specifically evaluated in the context of a well-designed prospective study, particularly in relation to maintenance of analgesia. Meanwhile, opioid rotation has been studied, albeit retrospectively, and an associated improvement in the “leading symptom‘ (mostly neurotoxic symptoms: cognitive impairment, myoclonus, and perceptual disturbance) and pain intensity was demonstrated in approximately 70% of patients.132 This study could not specifically evaluate the relative contributions of hydration and opioid rotation to the improvement. Another retrospective study in a different patient group noted a decrease of approximately 60% in the chart recordings of agitated impaired mental status in association with a practice change of increased opioid rotation, hydration, and objective cognitive monitoring.133 An open prospective study of hospice patients by Maddocks et al. found a significant improvement in mental state in patients with morphine-associated delirium, who were switched to oxycodone.38 In a recent prospective study of delirium in patients with advanced cancer, intervention in the form of opioid switching or dose reduction, and drug discontinuation or dose reduction in the case of other psychotropic medications, was found to be independently associated with delirium reversal.4 Retrospective and prospective studies of the occurrence rates, etiology, and management of cognitive dysfunction in this population are summarized in Table 7A, B respectively.

Table 7A. Retrospective Studies that Examined the Etiology or Management of Cognitive Dysfunction in Cancer Patients
ReferenceStudy population sampleOpioid doses and administrationCognitive dysfunctionManagement
AssessmentPrincipal findingsInterventionOutcome
  1. OR: opioid rotated or switched; ONR: opioid not rotated or switched; Mor: morphine; p.o.: orally; i.v.: intravenously; s.c.: subcutaneously; MEDD: morphine equivalent daily dose expressed in mg as mean ± standard deviation; MMSE: Minimental State Examination; AIMS: agitated impairment mental status; DSM-III-R: Diagnostic Statistics Manual 3rd ed. revised; CI: cognitive impairment.

Caraceni et al. (1994)131n = 15, case seriesMor p.o.; 20–160 mg (n = 12) Mor i.v. or s.c.: 30–82 (n = 3)MMSE, DSM-III-R criteria for deliriumMorphine was the sole etiologic factor (n = 4) Other contributory conditions likely (n = 11)Route switch (n = 1); dose reduction (n = 9)Delirium improved (n = 8) Authors concluded that opioids are seldom the sole cause of delirium
Bruera et al. (1995)1331988–89 year, n = 117; 1991–92 year, n = 162MEDD: 314 ± 596 (n = 117) MEDD: 308 ± 555 (n = 162)MMSE, recorded AIMSGreater than 50% reduction in the incidence AIMS in 1991–1992 period50% increase in hydration and opioid rotationReduced AIMS. Decreased neuroleptic and benzodiazepines use as treatment for agitation
De Stoutz et al. (1995)132n = 191 (total); OR (n = 80); ONR (n = 111)MEDD: 578 ± 1535 OR (n = 80) 250 ± 568 ONR (n = 111)MMSECognitive impairment (n = 42)Opioid switch in 42 patients with CI29 of 42 (69%) had an improvement in CI
Table 7B. Prospective Studies that Examined the Occurrence Rates, Etiology, or Management of Cognitive Dysfunction in Cancer Patients
ReferenceStudy population sampleOpioid useCognitive dysfunctionManagement
AssessmentPrincipal findingsInterventionOutcome
  1. M/F: males/females; Pr: prospective; Ret: retrospective; M: morphine; p.o.: orally; HM: hydromorphone; MEDD: morphine equivalent daily dose in mg; MMSE: Minimental State Examination, DSM-III: Diagnostic Statistics Manual 3rd ed.; CI: cognitive impairment; DSM-III-R: Diagnostic Statistics Manual 3rd ed. revised; SCID: structured clinical interview for DSM-III-R; MDAS: Memorial Delirium Assessment Scale; BOMC: Blessed Orientation Memory Concentration Test; CAM: Confusion Assessment Method; CRS: Confusion Rating Scale. M-3-G: morphine-3-glucuronide; M-6-G: morphine-6-glucuronide; AMS: altered mental status; OBS: organic brain syndrome.

Leipzig et al. (1987)40n = 35 consecutive admissions; M/F; 9/16; median age, 59 yrs (range 19–86); Controls, n = 8Mor, codeine, levorphanol, HM, oxycodoneModified MMSE27 of 35 (77%) had CI. Opioid-related sedation (n = 7). Organic brain syndrome (n = 8), 9 patients died, 8 with OBS at the time of deathOpioid dose reduced (n = 4); correction of various other factors (n = 11)Improvement in CI with dose reduction (n = 4) and correction of other factors (n = 11)
Folstein et al. (1983)393 prevalence samples: 1 day (n = 27), consecutive admissions (n = 83), 3 day (n = 50)Not statedMMSE1-day prevalence of 22%, prevalence rate of 26% in consecutive admissions, and 3-day prevalence of 14%Not studiedNot studied
Massie et al. (1983)6n = 19 chief oncology resident referrals to a psychiatric serviceNot statedDSM-III criteria for delirium13 patients were observed until death occurred. Multifactorial etiologyNot stated11 of 13 (85%) died in delirium
Bruera et al. (1992)5n = 61; mean age, 58 ± 12 yrs; M/F, 35/26Various opioids; doses not statedMMSE: 3 days/week; CI, a score <24 or a 30% decrease in score39 of 47 (83%) died with CI. Cause of CI unknown in 37 (56%). Most common causes: drugs, sepsis, and brain metastasesMultiple, as part of routine practice22/66 (33%) episodes of cognitive impairment improved
Maddocks et al. (1996)38n = 13; M/F, 7/6; median age, 73 (range 63–79) yrsMor p.o. 60 mg; Mor s.c.; 10–150 mg;DSM-IV criteria for deliriumDeveloped delirium (n = 13) Poor metabolizer of oxycodone (n = 1), extensive metabolisers (n = 12)Opioid switched to oxycodoneIn 8 of 12 (67%), delirium improved
Minagawa et al. (1996)44n = 93 patients admitted to a palliative care unitNot statedDSM-III-R delirium criteria, SCID, MMSE.Cognitive impairment occurred in 42%, delirium in 28%, and dementia in 10.7%Not studiedNot studied
Ashby et al. (1997)89n = 36 patients admitted to hospice; M/F, 18/18; mean age; 59 (24–79) yrsMEDD p.o.: 110 (20–600) mg (n = 17); MEDD s.c.: 50 (20–830) mg n = 19Confusion documented as present or absentSerum creatinine elevation correlated with adverse effects, which included delirium (n = 9). Dose corrected plasma M-3-G and M-6-G higher in patients with abnormal creatinine levels.Not studiedNot studied
Lawlor et al. (2000)4n = 104 admitted to a palliative care unitNot statedDSM-IV criteria for delirium, MMSE, MDASOccurrence rate of delirium was 42% on admission and 88% in the hours before deathMultiple,as part of routine practiceDelirium reversibility: 50% Reversal associated with psychoactive drugs and dehydration as precipitants
Tuma and DeAngelis (2000)43n = 140; median age, 73 (range, 26–92) yrs M/F, 69/71; Pr study: n = 100. Ret study; n = 40Not statedDSM-III-R delirium criteria. MMSE in Pr studied group.AMS on admission (n = 48; 34%). Developed AMS during admission (n = 92, 66%). Drugs, mainly opioids and benzodiazepines were the most common causes.Not studiedNot studied
Gagnon et al. (2000)42n = 89; mean age; 66.4 yrs; M/F: 43/46Median s.c. MEDD: 27.1 mgBOMC, CAM, CRS20% were delirious on admission. The incidence of delirium after admission was 32.8%Multiple, as part of routine practiceSignificant symptom improvement occurred in 50% of delirious patients

In the process of switching opioids in chronic use, physicians must cautiously interpret the standard equianalgesic tables, which were largely derived from single dose studies.134 In switching to the new opioid, a dose reduction of up to 50% of the standard equianalgesic dose has been recommended by some authors to account for incomplete cross-tolerance.135 Other authors have recommended a dose reduction of 50–75% of the equianalgesic dose ratio in the presence of good pain control and 0–25% reduction in the presence of poor pain control.52 These recommendations are made on the basis of clinical experience rather than systematic studies. In the case of methadone, the switch is best done on a phased basis by a physician experienced in its use, and acknowledging recent evidence of its underappreciated potency,136–138 which is not reflected in many of the standard equianalgesic tables.54, 135 Some authors have recommended a reduction to 25–33% of the equianalgesic dose of methadone on the basis of clinical experience.52 However, switching from other opioids to methadone using this approach still carries the risk of toxicity, depending on the ratio used, and especially in patients on higher doses. Evidence is now emerging that in switching to methadone there is a highly positive correlation between the equianalgesic dose ratio and the dose of prior opioid.136–139

Rather than switch opioid or reduce the opioid dose when inadequate pain control occurs in association with opioid toxicity, some authors suggest a change in the opioid administrative route to subcutaneous or spinal140 or intraventricular administration.141 A small randomized study of 10 patients showed that a switch from the oral route to either subcutaneous or epidural administration of morphine resulted in improved pain control and a reduction in adverse effects.142 There was no difference either in pain relief or in adverse effects between the subcutaneous and epidural routes. Route of administration can have a significant impact on opioid metabolite levels.88 In the case of orally administered morphine, both M-3-G and M-6-G plasma levels are higher than when morphine is administered parenterally.143, 144 A small case series examining the change from oral or subcutaneous to the intracerebroventricular route of morphine administration in cancer patients found that the switch was associated with nondetection of M-6-G and 90% reduction in M-3-G in cerebrospinal fluid samples.141 This coincided with improved pain control and diminution of side effects, including sedation. Use of the oral administrative route often becomes an impractical option in the presence of delirium. Although use of the subcutaneous route is feasible in most settings, the management of epidural or intraventricular administration presents significant practical challenges in many settings.

Is the propensity for opioid-related cognitive dysfunction less for some opioids? Two randomized, double blind, crossover studies comparing oxycodone and morphine suggest less perceptual disturbance occurs with oxycodone.145, 146 Clinical experience with the use of methadone suggests that this opioid is possibly associated with less opioid-related neurotoxicity. Clearly, methadone is more potent than currently quoted in many of the equianalgesic tables, and the equianalgesic dose ratio is not constant but has a positive correlation with the dose of other opioid before switching to methadone.136–138 The implication of the NMDA receptor in opioid tolerance,147 and the known NMDA receptor antagonist properties of methadone,148 could possibly explain the documented lesser escalation of opioid dose with methadone.149 The reduced need for dose escalation with methadone and the absence of known neurotoxic metabolites might result in less cognitive dysfunction with this opioid, although this has not been demonstrated to date.

A proposed approach to management is summarized in Table 8. Whatever the level of intensity applied to interventions directed at the underlying causes of delirium, symptomatic treatment of delirium is indicated in most cases, particularly where opioid toxicity may be associated with agitation. Haloperidol, a potent antidopaminergic neuroleptic is considered the drug of choice.126, 150 A randomized controlled trial in acquired immunodeficiency syndrome patients with delirium showed improvement in symptoms with either chlorpromazine or haloperidol but not with lorazepam.151 Methotrimeprazine given intravenously or subcutaneously is useful if a greater level of sedation is required, although postural hypotension is a recognized side effect.8 Midazolam is a short-acting benzodiazepine commonly used for sedation of refractory symptoms, particularly those of delirium, in the palliative care setting.152 This is given by intravenous or subcutaneous infusion and titrated to keep the patient sedated, particularly when treatment of the underlying causes is either not possible or ineffective, and delirium reversal therefore is not possible. In clinical practice the choice of neuroleptic or other sedating agent is largely determined by the degree of symptom distress (predominantly related to perceptual disturbance or delusional activity) and behavioral disturbance, especially agitation. In order of ascending magnitude of sedation, the choice of agent from among those previously discussed would likely be as follows: haloperidol, methotrimeprazine, and midazolam. A more comprehensive account of doses and side effects of these agents is beyond the scope of this article, and the reader is referred to more authoritative sources for these details.8, 153 In essence, there is a remarkable paucity of randomized controlled trials regarding sedating agents for symptomatic treatment of delirium.154 It is likely that some of the newer antipsychotic agents such as risperidone or olanzepine with fewer extrapyramidal side effects will be used to advantage in the treatment of delirium.155, 156 However, lack of controlled trials in this population and lack of parenteral formulations currently pose limitations to their use.

Table 8. Proposed Management Approach to Patients Presenting with Opioid-Related Delirium
• Educate family regarding nature of opioid-related cognitive dysfunction and delirium: psychomotor agitation, perceptual disturbance, delusions, and communication difficulties107–110
• Establish goals of care, evaluate pros and cons of intervention in the context of the patient’s clinical situation, previously expressed wishes, and family wishes128
• Monitor delirium severity with an instrument that is either partly118, 119 or wholly observational42, 122, 123
• Consider either opioid dose reduction or opioid switch129–133
• Discontinue or minimize use of all unnecessary medications, especially other psychotropics4
• Consider all other contributory causes, especially concurrent reversible causes4
• Assess renal function and adjust dose of opioid or other drugs accordingly92–97
• Assess hydration and supplement with parenteral fluids if necessary4
• Commence symptomatic treatment of delirium with neuroleptic such as haloperidol8, 150, 151
• Environmental manipulation: minimal noise, reorientation, encourage presence of family member116
• Prevent recurrence: if delirium is reversed, address risk factors for recurrence in future (see Table 6)

It is imperative that efforts are made to educate the patient's family regarding the nature of delirium, and in particular the potential for associated agitation and emotional lability to be misinterpreted as poor pain control.108, 109

The use of psychostimulants has been advocated in the treatment of opioid-induced sedation,157–159 and cognitive benefit has been demonstrated with neuropsychologic testing in a randomized, double-blind, placebo-controlled, crossover trial of methylphenidate in patients receiving a continuous infusion of opioids for cancer pain.160 The rationale behind the use of psychostimulants is that they counteract the sedation and probably also the cognitive impairment associated the upward titration of opioid. Methylphenidate doses of 10 mg at 8:00 a.m. and 15 mg at noon have been used with generally good effect in cancer patients with incidental pain.158 Clinical experience suggests that on initiating treatment, lower doses in the region of 2.5–5 mg are advisable in the elderly. Alternatives to methylphenidate include dextroamphetamine, pemoline, and modafinil. Pemoline is available as chewable tablets and can be absorbed sublingually. However, hepatotoxicity is an associated risk.161, 162 In summary, there are few data on the role of psychostimulants to counteract opioid-induced sedation, and further controlled trials are needed. Although psychostimulants have been proposed as a potential treatment for hypoactive–hypoalert delirium,163 there is clearly concern regarding the potential for aggravation of perceptual disturbance, which is not an uncommon occurrence in hypoactive delirium.164 However, in light of available evidence, psychostimulants would appear to be of benefit in selected patients, particularly in opioid-induced sedation, and in the absence of perceptual disturbance or delusional disorder.

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. GENERAL CONTEXT OF OPIOID-RELATED COGNITIVE DYSFUNCTION IN PATIENTS WITH CANCER
  4. PREVALENCE AND PATTERN OF COGNITIVE EFFECTS ASSOCIATED WITH OPIOID USE
  5. DELIRIUM AND COGNITIVE DYSFUNCTION: THE ROLE OF OPIOIDS
  6. PATHOPHYSIOLOGY OF OPIOID-ASSOCIATED COGNITIVE DYSFUNCTION
  7. IMPACT OF OPIOID-ASSOCIATED COGNITIVE EFFECTS ON PATIENT, FAMILY, AND STAFF
  8. ASSESSMENT OF COGNITIVE DYSFUNCTION IN OPIOID-TREATED PATIENTS
  9. MANAGEMENT APPROACH TO OPIOID-ASSOCIATED COGNITIVE IMPAIRMENT
  10. CONCLUSIONS
  11. REFERENCES

Cognitive dysfunction occurs frequently in the advanced cancer population, typically within the context of the syndrome of delirium. Its recognition warrants objective cognitive monitoring, which in itself is often reassuring to patients with undue fear of opioid-induced cognitive dysfunction. The opioid contribution to this dysfunction is sometimes difficult to evaluate in the presence of multiple other cancer disease and treatment-related factors, highlighting the need for a multidimensional approach to patient assessment. The presence of other features of opioid-induced neurotoxicity should prompt physicians to suspect an opioid contribution to the impairment. A management approach that addresses concurrent problems and either an opioid switch or dose reduction can lead to reversal of the cognitive impairment in up to 50% of episodes. Although opioid switching or dose reduction involves minimal burden or risk to the patient, the intensity of management intervention in addressing concurrent problems must be guided by the goals of care and the recognition that virtually all patients experience cognitive impairment in the days or hours before death.

Future research initiatives might further examine the phenomenology of cognitive dysfunction throughout the course of the illness, characterize the opioid contribution to the development of delirium, compare the efficacy of different neuroleptics, and develop predictive models for delirium reversibility.

REFERENCES

  1. Top of page
  2. Abstract
  3. GENERAL CONTEXT OF OPIOID-RELATED COGNITIVE DYSFUNCTION IN PATIENTS WITH CANCER
  4. PREVALENCE AND PATTERN OF COGNITIVE EFFECTS ASSOCIATED WITH OPIOID USE
  5. DELIRIUM AND COGNITIVE DYSFUNCTION: THE ROLE OF OPIOIDS
  6. PATHOPHYSIOLOGY OF OPIOID-ASSOCIATED COGNITIVE DYSFUNCTION
  7. IMPACT OF OPIOID-ASSOCIATED COGNITIVE EFFECTS ON PATIENT, FAMILY, AND STAFF
  8. ASSESSMENT OF COGNITIVE DYSFUNCTION IN OPIOID-TREATED PATIENTS
  9. MANAGEMENT APPROACH TO OPIOID-ASSOCIATED COGNITIVE IMPAIRMENT
  10. CONCLUSIONS
  11. REFERENCES