Impaired Neuropsychological Performance in Chronic Nonmalignant Pain Patients Receiving Long-Term Oral Opioid Therapy

100 Journal of Pain and Symptom Management Vol. 19 No. 2 February 2000 Original Article Impaired Neuropsychological Performance in Chronic Nonmali...
Author: Laurence Arnold
13 downloads 0 Views 152KB Size
100

Journal of Pain and Symptom Management

Vol. 19 No. 2 February 2000

Original Article

Impaired Neuropsychological Performance in Chronic Nonmalignant Pain Patients Receiving Long-Term Oral Opioid Therapy Per Sjøgren, MD, Annemarie B. Thomsen, MD, and Alf K. Olsen, MD Department of Palliative Medicine (P.S., A.K.O.), Bispebjerg Hospital, Copenhagen, and H:S Multidisciplinary Pain Center (A.B.T.), National Hospital, Copenhagen, Denmark

Abstract The study investigated neuropsychological performance in chronic nonmalignant pain patients receiving long-term oral opioid therapy. Forty patients treated solely with regular and stable doses of an oral opioid were compared with 40 healthy volunteers. The patients received daily opioid doses of 15–300 mg of oral morphine (median: 60 mg) or equianalgesic doses of other opioids. The neuropsychological tests consisted of continuous reaction time (CRT), which measured vigilance/attention; finger tapping test (FTT), which measured psychomotor speed; and paced auditory serial addition task (PASAT), which measured working memory. Three months after the study had been carried out, 14 of the controls were retested in order to determine the reliability of the three tests. The patients performed statistically significantly poorer than the controls in all the tests. Significantly positive correlations were found between the PASAT and pain visual analogue scales (VAS). In the retesting of 14 controls, it was found that the tests showed high reliability. Vigilance/attention, psychomotor speed, and working memory were significantly impaired in chronic nonmalignant pain patients. The present study cannot determine which factors influenced the test results, but pain itself seemed to have an arousal effect on working memory. J Pain Symptom Manage 2000;19:100–108. © U.S. Cancer Pain Relief Committee, 2000. Key Words Chronic nonmalignant pain, cognitive function, opioids, psychomotor function

Introduction In contrast to the well-established use of opioids for the treatment of acute pain and cancer pain, the administration of long-term opioid therapy for pain not due to malignancy remains controversial. However, a number of studies have described beneficial effects following long-

Address reprint requests to: Per Sjøgren, MD, Department of Palliative Medicine, Bispebjerg Hospital, Bispebjerg Bakke 23, DK-2400 Copenhagen NV, Denmark. Accepted for publication: March 24, 1999. © U.S. Cancer Pain Relief Committee, 2000 Published by Elsevier, New York, New York

term administration of opioids in selected patients suffering from pain of nonmalignant origin. The concerns about cognitive impairment, decreasing efficacy due to tolerance, and the development of drug dependence and addiction are the main reasons for the reluctance to prescribe long-term opioid medication in these patients.1,2 Patients with chronic nonmalignant pain referred to multidisciplinary pain clinics highlight these concerns and may present many problems, such as long treatment periods; a high proportion of time spent in hospitals; a large number of treating doctors; many failed 0885-3924/00/$–see front matter PII S0885-3924(99)00143-8

Vol. 19 No. 2 February 2000

Impaired Neuropsychological Performance

therapies, including surgical procedures; misuse of drugs, resulting in dependence and side effects; use of opioids at irregular intervals and/or by the parenteral route; and, finally, psychosocial comorbidities. Despite these concerns, an attempt at treatment with oral long-term opioid administration at regular intervals is sometimes undertaken when other therapeutic possibilities have been exhausted. In many Western countries, the use of opioids for long-term treatment of chronic nonmalignant pain is increasing in frequency.1,2 Denmark has the world’s highest legal consumption of opioids per capita and a recent investigation indicated that the majority of this consumption was prescribed for nonmalignant pain conditions.3 The literature concerned with cognitive and psychomotor functioning during long-term opioid treatment hitherto has primarily dealt with drug addicts and cancer patients,4,5 or short-term administration in healthy volunteers.6 Although patients with nonmalignant pain constitute a rather heterogeneous group, they are quite different from cancer patients in terms of pains, psychosocial circumstances, diseases, etc. To our knowledge, few clinical studies have attempted to examine cognitive and psychomotor functioning in chronic nonmalignant pain patients during long-term treatment with opioids.7 Recently, an interesting experimental study in chronic, nonmalignant pain patients has addressed the issue.8 Such studies in chronic pain of nonmalignant origin are critical for providing policy recommendations based on scientific data. The purpose of the present study was to compare some aspects of neuropsychological performance in chronic nonmalignant pain patients receiving stable, long-term oral opioid therapy with healthy controls.

Methods Patients Forty chronic nonmalignant pain patients (median age: 60 years, range: 46–74 years; 16 male, 24 female) receiving long-term oral opioid treatment at the Multidisciplinary Pain Center were investigated prospectively. Patients who were taking oral opioids on a fixed time schedule and who had had stable daily opioid doses ⬎ 14 days were eligible. Patients

101

taking other psychotropic drugs, those suffering from metabolic disturbances or drinking alcoholic beverages, and those with physical or neurological dysfunction interfering with the test did not participate in the study. Twenty-three patients received sustained-release morphine, 12 methadone, 2 ketobemidone, 2 buprenorphine, and 1 tramadol. All opioid doses were converted to equivalent milligrams of oral morphine; for opioids other than morphine, the equipotency table published by Clausen et al. was used for conversion.3 Daily opioid doses were 15–300 mg (median: 60 mg) of oral morphine or equianalgesic doses of other opioids. The last opioid dose had been ingested 50–450 min (median: 233 min) before testing. The patients receiving sustained-release morphine used a median 60 mg (range 20–300 mg) of morphine. The patients receiving methadone used a median equianalgesic dose of 60 mg (range 15–240 mg) morphine. The neuropsychological performance of the patients was compared with 40 healthy sex- and age-matched volunteers (median age: 59 years, range: 49–78; 11 male, 29 female). Three months after this investigation had been carried out, 14 of the healthy controls were retested in order to determine the reliability of the three tests. Healthy controls taking other psychotropic drugs, suffering from metabolic disturbances, drinking alcoholic beverages, or suffering from other physical or neurological dysfunction interfering with the tests did not participate in the study. The study was in accordance with the Helsinki Declaration and was approved by the Ethical Committee of Copenhagen. The patients and the controls gave informed and written consent.

Assessments All assessments were performed late in the morning. Pain localization, pain pathophysiology, opioids, and side effects were evaluated. Sedation and pain level were evaluated using 100 mm visual analogue scales (VAS) (sedation [SVAS]: “quite alert–extremely tired”; pain [PVAS]: “no pain–intolerable pain”). The neuropsychological tests consisted of continuous reaction time (CRT), finger tapping test (FTT), and paced auditory serial addition task (PASAT). All testing was performed in the above mentioned sequence and lasted approximately 1 hour. The neuropsychological

102

Sjøgren et al.

functions of attention/concentration (CRT), psychomotor speed (FTT), and working memory (PASAT) are considered particularly relevant parameters. These functions are considered to be “higher-order” cognitive functions not strictly localized to any single brain region. They are most crucial to general adaptive and information processing aspects of behavior.9 CRT is considered to measure vigilance, i.e., the ability of the individual to attend to and respond rapidly to external stimuli for an extended period of time. Vigilance tests examine the ability to sustain and focus attention in itself. The test can differentiate quantitatively and qualitatively between different types of cerebral dysfunction, whether induced pharmacologically or as a result of some other organic disturbance. Age, intelligence, and educational level have only a slight influence on CRT, and the learning effect is minimal.10 Through headphones, 152 auditory signals (500 Hz, 90 dB) were delivered to the patient at random intervals (2–5 sec) over a period of 10 min. The patients were instructed to press a button as soon as they heard the sound, using the first finger of the dominant hand. A computer registered the time from emission of the sound signal to activation of the button. Reaction times were measured in 1/100 sec. If the patient did not respond within 2 sec, this was recorded as “no response.” Reaction times ⬍12/100 sec indicate activation of the button without a preceding stimulus since nerve conduction velocity and delay in the apparatus make it impossible to respond faster. Almost all CRT distributions have the same skewed distribution with a thick tail to the right (positively skewed). According to Elsass,10 the distributions of the control groups and those cerebrally affected show so little overlap that there is no doubt that they are significantly different from each other. Instead of testing for statistical significance, investigations were made to see which CRT score gave the maximal separation between the groups (with the smallest amount of false positives and false negatives). By analyzing how a cutoff score for maximal classification, corrected for the number of observations, could differentiate, it was found that nonparametric percentiles separated better the moment scores. Furthermore the numbers of different forms of movement artifacts, and slow or fast CRTs, were analyzed

Vol. 19 No. 2 February 2000

for their capacity to differentiate cerebral affection. The percentile values, compared with other scores, made the best separation between groups, with the smallest numbers of false positives and false negatives. The discriminative capacity of the test using percentiles has formerly been described in studies of longterm opioid therapy.11,12 Accordingly, the CRT values were summarized using 10th%, 50th%, and 90th% percentiles. The 10th% percentile represents the fastest and the 90th% percentile the slowest values. FTT has been used in neuropsychology since the 1940s. In addition to being sensitive to lesions of motor structures of the cerebral hemispheres, detecting lateralized disability, psychomotor slowing is considered a nonspecific consequence of cerebral dysfunction. The test requires the patient to tap a key as fast as possible. The key was attached with a device for recording the number of taps. The second finger of each hand made five 10-second trials with brief rest periods between the trials. The score for each hand was the average for each five trials.13 PASAT measures working memory, another aspect of attention. The test requires addition of simple digits presented auditorily in several series of a successively higher pace of presentation. The task, which reflects the capacity for divided attention, is a measure of information processing speed and has appeared to be sensitive to minor attention deficits.14 The patient was verbally presented 61 random digits between 1 and 9 at timed intervals and was instructed continuously to add the last digit to the previous. The number of correct answers were counted. The test was repeated up to four times, increasing the speed. Initially the interval between each digit was 2.4 sec (T 2.4), then 2.0 sec (T 2.0), 1.6 sec (T 1.6), and finally 1.2 sec (T 1.2). PASAT thus yields an estimate of the subject’s ability to register sensory input, respond verbally, and retain and use a complex set of instructions. The patient must also hold each item after processing, retrieve the held item for addition to the next digit, and perform at an externally determined pace. Unfortunately, patients experience this sensitive test as very stressful and humiliating, which was the reason for only using the two longest intervals in the retest. The Hospital Anxiety and Depression Scale (HAD) was used to evaluate the influence of

Vol. 19 No. 2 February 2000

Impaired Neuropsychological Performance

103

Table 1 Retesting 14 Controlsa Tests CRT 10th percentile 50th percentile 90th percentile FTT DOM NDOM PASAT T 2.4 T 2.0

First assessment (mean SD)

Second assessment (mean SD)

Pb

rc

19.85 (6.08) 22.92 (7.02) 29.21 (8.51)

19.57 (5.52) 22.92 (6.29) 31.07 (12.41)

NS NS NS

0.67 0.86 0.36

50.80 (5.32) 44.44 (5.14)

51.87 (6.31) 45.90 (5.38)

NS NS

0.78 0.76

36.64 (12.36) 35.61 (9.92)

36.76 (11.39) 35.00 (9.08)

NS NS

0.83 0.96

a Mean and standard deviation (SD) for continuous reaction time (CRT), finger tapping test (FTT), and placed auditory serial addition task (PASAT) at the first and second assessments. b Significance probabilities of Wilcoxon signed rank sum test testing for within-patient difference between the first and second assessments. c Spearman’s rank correlation coefficient.

depression and anxiety on neuropsychological performance. HAD measures milder forms of depression and anxiety in medically ill patients. The norm is ⱕ16, e.g., maximum score of 8 on one of two subscales. A score above 8 in any of the subscales indicates clinically significant anxiety or depression.15 The Karnofsky Performance Status Scale was used to determine the patient’s physical function. The score was assessed by the treating physician and ranged from 0% to 100%. A score of 100% is “Normal. No complaints. No evidence of disease,” and 0% is “Dead”.16

was 39 (0–80 mm). The median Karnofsky Performance Status Scale was 70% (range: 60– 100%). In the healthy control group, the median SVAS was 4 mm (range: 0–54 mm) and the median PVAS was 0 mm (range: 0–10 mm). The Karnofsky Performance Status Scale was 100% in all the controls. The HAD scores indi-

Statistics

Pain localizations Back Extremities Thoracic Neck Abdominal Total body Head or face Pain pathophysiology Somatic Neuropathic Visceral Opioids Sustained-release morphine Methadone Others (2 ketobemidone, 2 buprenorphine, and 1 tramadole) Side effects Constipation Dry mouth Sedation Dizziness Itching Sweating Nausea Hallucinations

Nonparametric statistical analyses were applied; mean values and standard deviations (SD) are presented in Table 1. The MannWhitney rank sum test was used for betweengroup analysis (unpaired data). The Wilcoxon signed rank test was used for within-group analysis (paired data). Correlations were analyzed by means of Spearman’s rank test, rs. The general level of significance was set at P ⫽ 0.05. The number of patients included in the study was driven from statistical power calculations based on a former study12 and a statistical analysis when half of the patients were included.

Results Pain localization, pain pathophysiology, opioids, and side effects are shown in Table 2. In the patient group, the median SVAS was 27 mm (range: 0–78 mm) and the median PVAS

Table 2 Pain Localization, Pain Pathophysiology, Opioids, and Side Effects Number

%

18 7 5 3 3 3 1

45.0 17.5 12.5 7.5 7.5 7.5 2.5

30 18 4

75.0 45.0 10.0

23 12

57.5 30.0

5

12.5

26 24 21 11 7 6 5 2

65 60 53 28 18 15 13 5

104

Sjøgren et al.

Vol. 19 No. 2 February 2000

cated that 50% of the patients suffered from anxiety (median: 11; range: 8–17) and 38% had depression (median: 10; range: 8–16). The CRT values (Figure 1) were summarized using 10th, 50th, and 90th% percentiles. The 90th% percentile showed statistically significantly slower values in the patient group (P ⫽ 0.022). On FTT (Figure 2), the patients performed statistically significantly slower than the controls using dominant and nondominant hands (P ⫽ 0.008 and P ⫽ 0.003, respectively). On the PASAT, statistically significantly poorer results were obtained by the patients at the longest three intervals (T 2.4 sec: P ⫽ 0.005; T 2.0 sec: P ⫽ 0.014; T 1.6 sec: P ⫽ 0.02) (Figure 3). Patients receiving methadone (n ⫽ 12) were compared with patients treated with sustainedrelease morphine (n ⫽ 23).The patients receiving methadone performed statistically significantly slower in the FTT using the dominant hand (P ⫽ 0.05), but no other differences were found.

Fig. 2. Finger tapping test (FTT). Medians, ranges and 25 and 75% percentiles for numbers of taps with dominant (DOM) and nondominant hands (NDOM) in the patient group (P, N, DOM ⫽ 38, N, NDOM ⫽ 36) and control group (C, n ⫽ 40). Patients performed statistically significantly slower using both dominant (P ⫽ 0.008) and nondominant hands (P ⫽ 0.003).

Correlations Between Tests and Other Variables

Fig. 1. Continuous reaction time (CRT). Medians, ranges, and 25 and 75% percentiles within the 10th, 50th, and 90th% percentiles in the control group (C, n ⫽ 40) and the patient group (P, n ⫽ 40). The 10th% percentile represents the fastest and the 90th% percentile the slowest values. The 90th% percentile showed statistically significantly slower values in the patient group (P ⫽ 0.022).

In the patient group, only a few statistically significant correlations were found among SVAS, PVAS, HAD, Karnofsky Performance Status Scale, opioid doses, time from ingestion of last opioid dose to testing, side effects, and performance of the three neuropsychological tests. CRT 50th% percentile correlated with Karnofsky Performance Scale (P ⫽ 0.040, r ⫽ ⫺0.31). FTT for dominant hand correlated with depression (P ⫽ 0.049, r ⫽ 0.33) and age (P ⫽ 0.009, r ⫽ ⫺0.42). PASAT T 2.4 sec and T 2.0 sec correlated with PVAS (P ⫽ 0.025, r ⫽ 0.420 and P ⫽ 0.017, r ⫽ 0.601, respectively).

Intertest Correlations Statistically significant correlations between the neuropsychological tests could also be

Vol. 19 No. 2 February 2000

Impaired Neuropsychological Performance

105

hypothesis of concordance between the two assessments could not be rejected (P ⬎ 0.05). Because of the reasons mentioned earlier, only the longest two intervals of PASAT (T 2.4 sec and T 2.0 sec) were used. The concordance between the two test performances is numerically illustrated by generally high Spearman correlations (average r ⫽ 0.8). Thus the tests showed high reliability (Table 1).

Discussion

Fig. 3. Paced auditory serial addition task (PASAT). Medians, ranges, and 25 and 75% percentiles for numbers of correct answers in the control (C) and patient (P) groups. At the longest intervals (T 2.4 sec) 29 patients and 37 controls completed the test. Fifteen patients and 32 controls completed the interval: T 2.0 sec. Thirteen patients and 31 controls completed the interval: T 1.6 sec. Only 11 patients and 27 controls passed the last interval: T 1.2 sec (P ⫽ 0.25). The first two intervals are shown in the figure. At the first three intervals statistically significantly poorer results were obtained by the patients (T 2.4 sec: P ⫽ 0.005; T 2.0 sec: P ⫽ 0.014; T 1.6 sec: P ⫽ 0.02).

demonstrated in the control group CRT 90th% percentile correlated with PASAT T 2.4 sec and T 2.0 sec (P ⫽ 0.030, r ⫽ ⫺0.36 and P ⫽ 0.037, r ⫽ ⫺0.37, respectively). CRT 50th% percentile correlated with PASAT T 2.0 sec (P ⫽ 0.041, r ⫽ ⫺0.36). Finally, CRT 10th% percentile correlated with FTT for the dominant hand (P ⫽ 0.044, r ⫽ ⫺0.32), and CRT 10th% and 50th% percentiles correlated with FTT for nondominant hand (P ⫽ 0.022, r ⫽ ⫺0.36 and P ⫽ 0.024, r ⫽ ⫺0.36, respectively).

Reliability Test On retesting (CRT, PASAT, and FTT) the 14 controls 3 months later, it was found that the

The neuropsychological tests used in the present study are all well validated and give information about arousal and vigilance, the ability to concentrate on a task, the ability to divide attention, the fluency of motor sequences, the coordination of perception and movements, motor performance, and short-term memory. The tests were selected because they are considered to represent cognitive and psychomotor functioning most crucial to general adaptive and information processing aspects of behavior.9 It seems reasonable to assume that a “normal” performance in the tests is prerequisite for optimal performance of tasks of everyday life, e.g., driving ability, operation of machinery, etc. Furthermore “normal” performance of the tests may well be associated with other aspects of “quality of life” demanding vigilance, ability to concentrate, motivation, attention, and intact memory. This study design permitted variation in opioids and doses, recorded time from last medication, Karnofsky’s Performance Status Scale, pain intensity, sedation, and other variables. By choosing this design, information is gained reflecting the actual performance of this group of patients, rather than isolated opioid effects. The goals of therapy in chronic nonmalignant pain patients are a reduction of symptoms, suffering, and environmental reinforcers of pain behavior, as well as functional improvement, less dependence on the health care system, and, if possible, return to work. The usual treatment approach is multimodality, including physical exercise programs, encouragement of activity, and cognitive, behavioral, and/or supportive measures to provide functional restoration despite continuing pain.17 Analgesics are only part of this treatment. A powerful opioid drug may alleviate pain and improve rehabilitation, but may also encour-

106

Sjøgren et al.

age passivity and make patients less compliant with their treatment program and daily life.18 The patients presented in this study had suffered from pain for many years, and in general they had used opioids long before they were referred to the pain center. They had not been successfully controlled with other interventions, and one of the treatment goals was to convert and stabilize uncontrolled, on-demand use of short-acting opioids into a stable, controlled and regular administration of long-acting opioids.19,20 Long-term opioid treatment of chronic nonmalignant pain patients is still controversial. With regard to efficacy, some authors are rather enthusiastic, but others are more cautious.1 Concerning “traditional” opioid side effects, some reports indicate that they, quantitatively and qualitatively, are comparable to those complicating the treatment of cancer pain.1,21 These findings are in accordance with our study (Table 2). Studies of cognitive and psychomotor functioning in cancer patients treated with opioids began to be published almost a decade ago.11,22 Cognitive impairment in cancer patients is frequent and ascribed to a variety of causes, which can be classified broadly according to etiology: disease-related and treatment-induced factors. The neuropsychological effect of long-term opioid administration in cancer patients is a matter of interest. While some studies suggest that opioids cause psychomotor and cognitive dysfunctioning, other studies fail to demonstrate such impairments or even claim that opioids have beneficial effects on psychomotor functioning and cognition. In papers studying reaction time in cancer patients treated for pain, those on opioid analgesics had significant retardation in reaction time when compared to healthy controls11,12 or to cancer patients not on opioids.23 Clemons et al.24 examined advanced cancer patients on stable oral morphine and compared them to advanced cancer patients not on opioids and to healthy age-matched controls. Cancer patients performed less well than healthy controls in all assessments, and those treated with morphine had poorer grammatical reasoning, alertness, and cognitive functioning than both other groups. Cancer patients on regular morphine undergoing a series of psychological and neurological tests for assessment of driving ability

Vol. 19 No. 2 February 2000

were compared to those not on opioids. No significant difference was noted in driving ability, but those patients on morphine tended to perform less well. The authors concluded that cancer patients receiving long-term analgesic medication with stable doses of morphine do not have psychomotor effects of a kind that would be clearly hazardous in traffic.5 Thus, the sparse literature concerning cancer patients is conflicting. Studies have emerged investigating psychoactive drugs, for example, amphetamines, that can counteract opioid-induced sedation and cognitive dysfunction and possibly also the fatigue caused by the cancer disease itself.25 Given the increasing use of long-term opioid therapy in nonmalignant pain, further study of this problem is warranted. Our study showed significantly reduced PASAT, CRT, and FTT scores, as well as typical opioid side effects in the patient group. By choosing this design, information is gained on the actual performance of this group of patients. It is not possible, however, to determine whether opioids per se are responsible for the poorer neuropsychological performance. It may be that this group of patients is even more difficult than cancer patients to control for pain, performance status, and other symptoms. As a result, the psychosocial impact on neuropsychologic functioning may be pronounced. In this study, a substantial number of patients indicated anxiety (50%) and depression (38%). This also may influence neuropsychological performance, although no correlation with the tests could be demonstrated. An effect size analysis of neuropsychological functioning in patients with major depressive disorder using meta-analytic principles was recently conducted. Although other assessments were used, the results from 726 patients with depression and 795 healthy normal controls revealed that depression had the largest effect on measures of encoding and retrieval from episodic memory. Intermediate effect sizes were recorded on tests of psychomotor speed and tests that require sustained attention. Minimal effect sizes were found on tests of semantic memory, primary memory, and working memory.26 The patients in this study also suffered from significant disability. The median Karnofsky Performance Status Scale was 70%, which means “Cares for self. Unable to carry on nor-

Vol. 19 No. 2 February 2000

Impaired Neuropsychological Performance

mal activity or to do active work.” The Karnofsky Performance Status Scale was originally designed for and has been widely used with cancer patients but has also been used with nonmalignant pain patients in chronic opioid therapy.1 The scale may be useful in this context, as it emphasizes physical performance and dependency, but the lower end of the scale obviously aims at terminally ill cancer patients.16 Significant correlations between PASAT T 2.4 sec and especially T 2.0 sec, and PVAS indicate that pain itself may have an arousal effect that improves neuropsychological functioning. On the other hand, severe pain may well lessen the ability to concentrate.5 A recent experimental study in chronic nonmalignant pain patients supports the notion that pain itself has an impact on cognition.8 Six opioid-naive patients in chronic pain were treated with sustained-release morphine and dose-increased until adequate pain relief was obtained. They were tested before treatment and after 7–14 days with subjective, behavioral, and neurophysiologic measures. The results indicated that perceptual–cognitive status was improved, probably due to the removal of pain as a mental stressor. The patients’ PVAS improved from 73 mm ⫾ 23 mm (mean ⫾ SD) to 27 mm ⫾ 21 mm (mean ⫾ SD). Although the patients in this study8 may be quite different from those in the present study, some mechanisms may be common. Pain may influence cognition to a higher degree than opioids. In the present study, the median PVAS was 39 (range: 0–80 mm), and the presence of some residual pain may have prevented opioid sedation. There may be a mutual compensation of neuropsychological impairment by opioids and pain at an adequate titration level.

Conclusion Vigilance/attention, psychomotor speed, and working memory were significantly impaired in chronic nonmalignant pain patients receiving long-term oral opioid therapy. Factors such as pain, performance status, and mood may influence the results. Positive correlations between the measurement of working memory and pain scores indicated that pain itself may have an arousal effect. Further studies are needed in order to evaluate and isolate the influence of these factors on neuropsychologi-

107

cal functioning in chronic nonmalignant pain patients.

References 1. Zenz M, Strumpf M, Tryba M. Long-term oral opioid therapy in patients with chronic non-malignant pain. J Pain Symptom Manage 1992;7:69–77. 2. Portenoy RK. Chronic opioid therapy in nonmalignant pain. J Pain Symptom Manage 1990;5:46– 52. 3. Clausen TG, Eriksen J, Molke F. Legal opioid consumption in Denmark: 1981–1993. Eur J Clin Pharmacol 1995;48:321–325. 4. Rothenberg S, Schoffenfield S, Meyer RF, Kraus B, Gross K. Performance differences between addicts and non-addicts. Psychopharmacology 1977; 52:299–306. 5. Vainio A, Ollila J, Matikainen E, Rosenberg P, Kalso E. Driving ability in cancer patients receiving long-term morphine analgesia. Lancet 1995;346: 667–670. 6. Kerr B, Hill H, Coda B, Calogero M, Chapman CR, Hunt E, Buffington V, Mackie A. Concentration-related effects of morphine on cognition and motor control in human subjects. Neuropsychopharmacology 1991;5:157–166. 7. Haythornthwaite JA, Menefee LA, QuatranoPiacentini AL, Pappagallo M. Outcome of chronic opioid therapy for non-cancer pain. J Pain Symptom Manage 1998;15:185–194. 8. Lorenz J, Beck H, Bromm B. Cognitive performance, mood and experimental pain before and during morphine-induced analgesia in patients with chronic non-malignant pain. Pain 1997;73:369–375. 9. Lezak MD. Neuropsychological assessment, 3rd ed. New York, Oxford: Oxford University Press, 1995:352–793. 10. Elsass P. Continuous reaction times in cerebral dysfunction. Acta Neurol Scand 1986;73:1–22. 11. Sjøgren P, Banning A. Pain, sedation and reaction time during long-term treatment of cancer patients with oral and epidural opioids. Pain 1989;39: 5–11. 12. Banning A, Sjøgren P. Cerebral effects of longterm oral opioids in cancer patients measured by continuous reaction time. Clin J Pain 1990;6:91–95. 13. Peters M. Prolonged practice of a simple motor task by preferred and non-preferred hands. Percept Mot Skills 1976;42:447–450. 14. Gronwall D, Wrightson P. Delayed recovery of intellectual function after minor head injury. Lancet 1974;2:995–997. 15. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiat Scand 1983;67: 361–370.

108

Sjøgren et al.

16. Karnofsky DA, Abelman WH, Craver LF, Burchenal JH. The use of the nitrogen mustards in the palliative treatment of carcinoma. Cancer 1948;1: 634–656. 17. Turk C, Meichenbaum DA. A cognitive–behavioural approach to pain management. In: Wall P D, Melzach R, eds. Textbook of pain. Edinburgh: Churchill Livingstone, 1994:1337–1348. 18. Moulin DE, Jezzi A, Amireh R, et al. Randomized trial of oral morphine for chronic non-cancer pain. Lancet 1996;347:143–147. 19. Savage SR. Long-term opioid therapy: assessment of consequences and risks. J Pain Symptom Manage 1996;11:274–286. 20. Becker N, Thomsen AB, Olsen AK, Sjøgren P, Bech P, Eriksen J. Pain epidemiology and health related quality of life in chronic non-malignant pain patients referred to a Danish multidisciplinary pain centre. Pain 1997;73:393–400. 21. Schofferman J. Long-term use of opioid analgesics for the treatment of chronic pain and non-ma-

Vol. 19 No. 2 February 2000

lignant origin. J Pain Symptom Manage 1993;8:279– 288. 22. Bruera E, Macmillan K, Hanson J, MacDonald N. The cognitive effects of the administration of narcotic analgesics in patients with cancer pain. Pain 1989;39:13–16. 23. Banning A, Sjøgren P, Kaiser F. Reaction time in cancer patients receiving peripherally acting analgesics alone or in combination with opioids. Acta Anaesthesiol Scand 1992;36:480–482. 24. Clemons M, Regnard C, Appleton T. Alertness, cognition and morphine in patients with advanced cancer. Cancer Treat Reviews 1996;22:451–468. 25. Bruera E, Fainsinger R, MacEackern T, Hanson J. The use of methylphenidate in patients with incident cancer pain receiving regular opiates: a preliminary report. Pain 1992;50:75–77. 26. Zakzanis KK, Leach L, Kaplan E. On the nature and pattern of neurocognitive function in major depressive disorder. Neuropsychiatry, Neuropsychology, and Behavioral Neurology 1998;11:111–119.

Suggest Documents