| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
From the *Department of Anesthesiology, University of Cologne, Cologne, Germany; Department of Palliative Care, University of Aachen, Aachen, Germany; and Pain Clinic, Department of Anesthesiology, University of Dresden, Germany.
Address correspondence to Rainer Sabatowski, MD, Pain Clinic, Department of Anesthesiology, University of Dresden, Fetscherstr. 74, 01307 Dresden, Germany. Address e-mail to rainer.sabatowski{at}uniklinikum-dresden.de.
Abstract
BACKGROUND: The therapeutic use of opioids has been associated with altered cognition and impaired psychomotor function. Several studies have demonstrated the impact of opioid therapy on psychomotor performance and cognition, but there are no data about the effect of long-term treatment with transdermal buprenorphine on driving ability.
METHODS: Thirty patients suffering from chronic noncancer pain, who had been treated with stable doses of transdermal buprenorphine, included in a prospective trial and compared with 90 healthy volunteers (matched pairs). A computerized test battery, developed to assess the driving ability of traffic delinquents in Germany, was used. Attention reaction, visual orientation, motor coordination, and vigilance were evaluated. The data from 14 variables were assessed, and for each test, a relevant score was defined. As the primary end-point, the sum score of the three relevant scores was determined. A weaker statistical means to assess the patient's performance is to compare the test results to an age-independent control group. Individuals performing worse than the 16th percentile of this control group are considered to be unable to drive according to German law.
RESULTS: According to tests that predict driving ability, patients receiving transdermal buprenorphine were shown to be noninferior to the control group. Driving ability, as defined as a result above the 16th percentile, did not differ significantly between the patients and the control group.
CONCLUSION: Long-term use of transdermal buprenorphine for chronic noncancer pain does not impair driving ability, but because of the individual variability of test results, an individual assessment is recommended.
Buprenorphine, a partial µ-opioid receptor agonist and an antagonist of the 
-opioid receptor, was introduced in Europe 20 yr ago. The transdermal formulation of buprenorphine (Transtec®, Grünenthal, Aachen, Germany) as a matrix patch was first launched in Germany and Switzerland in 2001 and is now being prescribed in Europe and Australia for chronic noncancer and cancer pain treatment (1–4).
There is an increasing acceptance of long-term opioid therapy for the treatment of chronic noncancer pain (5–8). However, the evidence whether opioid therapy can be associated with cognitive impairment or reduced psychomotor function is inconclusive (9–12).
Since driving ability is considered an important aspect of self-determination, the effect of opioids on complex psychomotor and cognitive function is of interest especially for patients with noncancer receiving long-term opioid therapy (13–17). Clinical experience and previous studies have shown that most opioid-related side effects decrease under stable long-term opioid use (18–20), but there are little valid data explicitly assessing driving ability (11,13–17,21,22).
Driving ability can be compromised if one or more of the different aspects of cognitive and psychomotor function, such as attention, reaction time, visual orientation, perception, vigilance, and motor coordination, are impaired (13,16). To assess all these factors, a test battery and statistic analyses complying with international recommendations as well as German law were performed (23–28).
We studied patients receiving a stable dose of transdermal buprenorphine (TDB) for the treatment of chronic noncancer pain to assess the impact of TDB on tests that predict driving ability.
METHODS
This was a prospective comparison of patients treated with TDB for chronic noncancer pain and a historical control group of healthy volunteers. Subjects and controls were matched for age and sex, with three controls selected for every patient with TDB.
Patients
Male and female outpatients older than 18 yr suffering from noncancer pain responsive to opioids could be enrolled if they had been treated with TDB for at least 4 wk and without a dose change in the previous 12 days. Participants also were required to have a valid driving license and the ability to speak and write German. Patients were excluded from the study if they were receiving benzodiazepines or barbiturates >3 times per week, high-dose antidepressant treatment (e.g., >75 mg amitryptiline per day) or regular antihistamines. Patients with physical disabilities, severe psychiatric or neurological diseases, or visual disorders that would prevent them from performing the tests were also excluded. All patients gave written, informed consent to participate. The study protocol and the consent form were approved by the ethics committee of the University of Cologne.
Control Group (CG)
Controls were randomly selected from a pool of healthy volunteers who had been tested between March 1996 and March 1998 (between 2 and 5 pm) at the Institute for Traffic Safety of the Technischer Überwachungsverein GmbH in Cologne, Germany. This pool was part of a larger sample composed of healthy volunteers, with five men and five women for each year of age from 18 to 80 yr. The control sample has been described as representative of the normal German population with regard to activity, autonomy, and driving experience (25,29).
Course of the Study
After patients had been informed about the study and given their consent to take part, their personal details (age, gender, etc.) and medical history were recorded, including full details of their pain disease and the treatments they were receiving. Participants were also asked about their driving experience. Testing was performed between 1 and 3 pm within 1 wk after screening. Before testing, a urine sample was taken to screen for the use of drugs not reported by the patients in the pain clinic. Data from patients using unreported drugs were included in the intention-to-treat (ITT) analysis whereas the remaining patients were analyzed as the per-protocol (PP) group. Pain intensity was rated immediately before testing using an 11-point numerical rating scale (NRS) ranging from 0 (no pain) to 10 (worst pain that can be imagined).
Test Battery
The test battery followed the German national recommendations on tests to determine driving ability (30). These require assessment of: performance under pressure, orientation, concentration, attention, and reaction time. Three of the tests performed in the present study (cognitive (COG), determination test (DT) and test for visual orientation, tachistoscopic perception (TAVT), described subsequently) covered these five areas and were defined as the relevant tests before the examination (24). Test batteries similar to the one used in this study are used for traffic delinquents in Germany. Permission to drive for traffic delinquents is usually denied if one or more of these tests is failed, i.e., if the test result is below the 16%-percentile of the age-independent reference range (data of a group of young healthy volunteers) (26). In addition, previously validated tests for motor coordination and for vigilance were also used in this study.
All tests were performed under standardized conditions with standardized instructions and in the same sequence by means of the computerized test system ("Vienna Test System"). Raw data were measured as well as combined scores. The entire test battery normally takes about 75 min to perform, with the vigilance test at the end taking 25 min.
Attention Test (COG)
Four pictures (numbers, letters, figures, etc.) were presented in a row with another picture below. Subjects had to decide whether the lower picture matched any of the four pictures above. A new set of pictures was presented either after a response or automatically after 1.8 s. Up to 200 sets of pictures were used in this test. The number of correct and incorrect responses, and the mean time to a correct response, mean reaction time (MRT) were recorded. The score was calculated as the sum of MRT and the square root of the product of MRT and mistakes (31).
Test for Reaction Time Under Pressure (DT)
Subjects were given a series of different audiovisual signals. Color symbols presented on the screen and acoustic signals had to be answered by corresponding buttons on the panel, symbols on the right or left side of the screen by corresponding pedals. The frequency of the stimuli was automatically adapted to the subject's response. This test took 480 s and the mean time to a correct response (MRT) was used as the score (32).
Test for Visual Orientation, Tachistoscopic Perception
A complex picture of a situation commonly encountered in traffic was presented for 0.8 s. Subjects had to decide whether the picture showed: pedestrians, cars, bicycles, traffic signs, and/or traffic lights (corrections were possible). A total of 33 situations were presented. The number of missed or wrongly identified elements was used as the score (25).
Test for Motor Coordination (2-Hand)
Subjects had to keep a signal on a track by turning two steering wheels: one controlling horizontal movements, the other vertical movements. The track consisted of three different sections (circle, V-shape, and L-shape) and had to be negotiated 19 times. The mean time taken to negotiate the track (T, in seconds) and the mean percentage of total time during which the signal was off the track (Off%) were recorded. The score was calculated as (T x Off%)/100 + 0.1 x T (33).
Vigilance Test (VIG)
Subjects were presented with a circle consisting of separate small spots on a monitor. A bright spot moved stepwise around this circle, like the hand of a watch. At long, but irregular intervals, the spot sometimes missed one of the positions (i.e., jumped over the marker spots). When this occurred the subjects had to press a button as fast as possible. The number of mistakes (incorrect responses or undetected jumps) and the mean time to a correct response (MRT) were recorded. The score was calculated as the sum of MRT and the square root of the product of MRT and mistakes (34).
Passed Tests
Another method of evaluating driving ability from the cognitive tests used here is to assume unimpaired driving ability if all test results are above the 16%-percentile of the age-independent reference range (26).
Statistical Methods
The study was designed as a noninferiority trial, i.e., the object was to demonstrate that patients treated with TDB did not perform significantly worse in the tests than the untreated controls. That means that their performance is not inferior when compared with that of the CG.
In such trials, a clinically significant difference (
, ) has to be defined. Alcohol has been used as a standard to assess the degree of impairment induced by several drugs (35). A blood alcohol level of 
0.05% has been shown to cause a marked impairment of driving ability and is the threshold for being unfit to drive under German law (36–38). In a previous study, the effect of different antidepressants on cognitive and psychomotor function was compared using a computerized test battery similar to our study. During this study, patients received alcohol orally with a targeted blood concentration of 0.05%. The strongest impairment was seen in the testing of vigilance (38). From the data of that study, an effect size of = 0.57 for the alcohol-related impairment of vigilance was calculated. Using this effect size, the raw values of the CG in our study were transformed to obtain virtual values that would be equivalent to test performance under the influence of 0.05% blood alcohol.
Using this assumption, noninferiority in the test battery results of the patients under opioid compared with those of controls can be interpreted as a performance significantly better than that of the CG with a blood alcohol concentration of 0.05%. The sample size needed to demonstrate noninferiority using 1:1 randomization was calculated as 39 (one-sided t-test, 
= 0.05, ß = 0.20) assuming no difference between patients and controls. To reduce the required number of patients, we decided to perform a 1:3 randomization, i.e., three controls were matched to each patient. This gave a sample size of 26 patients and 78 controls. We therefore aimed to enroll 30 patients to allow for drop-outs or protocol violators.
Each of the five tests used involved the recording of several parameters. To reduce the problem of multiple testing, one "relevant score" was defined before the study. The primary end-point was defined as the sum of the scores of the DT, COG, and TAVT tests after z-transformation of the individual scores using the mean and the standard deviation of the whole sample (39). Testing was performed using the Mann–Whitney U-test. A one-sided P value <0.05 was regarded as significant. Significance tests for parameters other than the primary end-point are exploratory in nature and were performed without adjustment for . Unless stated otherwise, results are presented as arithmetic mean ± sd and P values correspond to the test for noninferiority as described above.
RESULTS
Between October 2003 and November 2005, 30 outpatients were enrolled and matched to 90 controls. As a result of matching, the study and control populations had similar demographic characteristics (Table 1). The most frequent diagnosis in the buprenorphine group was lower back pain (n = 17) and neuropathic pain (n = 8; e.g., postherpetic neuralgia, plexopathia) or pain from miscellaneous diseases (n = 5) such as osteoporosis or arthritis (Table 1). The median duration of pain was 62 months (mean, 99; range, 2–400). Mean current pain intensity was rated as 4.2 ± 2.9 measured on an 11-step numerical rating scale (0: no pain, 10: worst pain that can be imagined). The patients had been treated with TDB for an average of 52 days (range 14–271) before testing and received an average dose of 45 ± 20 µg/h (range: 17.5–105 µg/h). Six patients felt incapable of driving a car when the test was performed.
|
Urine screening detected the unreported use of amitriptyline in two patients and doxylamine in one patient. Other substances, such as barbiturates, were not detected. Data from these three patients were included in the ITT group, whereas the remaining 27 patients were analyzed as the PP group. The results of the tests are displayed in Table 2.
|
Determination Test
The number of "processed items" and the "number of wrong reactions" were not available for the CG. The number of "correct reactions" of the buprenorphine group was less than those of the CG, but the difference was not statistically significant. The ITT group performed better than the PP. The reactions times were shorter than those of the CG (+) but a significant noninferiority could not be demonstrated for any of the parameters or the test score (Table 2).
Cognitive
Because of incomplete testing, only 29 patients could be assessed. The number of "right answers" was higher in the CG and the number of "wrong answers" was higher in the buprenorphine group. Compared with the CG + , the buprenorphine group performed better, but the differences were not statistically significant neither for the ITT nor for the PP. The MRT was similar in the buprenorphine and the CG + groups (Table 2).
Test for Visual Orientation, Tachistoscopic Perception
No data were available about the "processing time" of the CG. Because of incomplete testing only, 29 patients could be evaluated. Significant noninferiority could be shown in the analysis of "wrong answers" and the test score for ITT and PP compared with CG + (P < 0.01) (Table 2).
2-Hand
Because of incomplete testing, only 29 patients could be evaluated. The average time needed to complete the test was lowest for the ITT, followed by the PP group and the CG. Significant noninferiority could be shown in both analyses (PP: P < 0.01; ITT: P < 0.01). Analysis of the percentage of "time off track" as well as the calculated score had shown significant noninferiority for the ITT- and PP-group (P < 0.01) (Table 2).
Vigilance Test
The average number of "wrong answers" was lowest in the PP group, followed by ITT and CG. Significant noninferiority could be shown in both analyses (PP: P < 0.01; ITT: P < 0.05). The difference in the MRT for ITT, PP groups proved to be significantly noninferior (PP: P < 0.05; ITT: P < 0.01). Analysis of the calculated score demonstrated significant noninferiority only for the PP group (P < 0.05) (Table 2).
Sum Score (Primary End-Point)
For the sum score of the z-transformed DT-, COG-, and TAVT-scores, representing the cognitive items of the test battery, significant noninferiority could be demonstrated for the ITT group in comparison to the CG + (P = 0.01) and for the PP group in comparison to the CG + (P = 0.018) (Fig. 1).
|
Passed Tests
The percentages of patients who passed the single tests, that is, whose relevant test score was above the 16th percentile, are displayed in Figure 2. The subjects from the CG passed an average of 4.3 (sd = 1.1) of the five tests, similar to the patients from the ITT group with 4.3 (sd = 1.0) passed tests. The percentage of subjects passing all five tests was 63% for the CG and 58% for the ITT group. In COG, 2-Hand, and VIG tests, fewer patients from the ITT group reached the 16th percentile compared with those from the CG. The percentage of patients passing the test was higher for groups ITT in TAVT and the same in the DT compared with the CG. None of these differences was statistically significant.
|
To assess the influence, factors, such as buprenorphine dosage, current pain intensity, age, and driving experience, were correlated to the test results using a multivariate analysis. The buprenorphine dosage correlated with the number of wrong answers in VIG (r = 0.48; P = 0.027). Age correlated with the MRT in DT (r = 0.441; P = 0.015), the number of wrong answers (r = 0.449; P = 0.015), and the score (r = 0.447; P = 0.015) in TAVT. Age correlated negatively with the number of correct answers in DT (r = –0.492; P = 0.006) and the percentile reached in TAVT (r = –0.588; P = 0.001). No correlation was found between current pain intensity and driving experience and any of the examined parameters.
The test results of the six patients who felt incapable of driving a car when the test was performed did not differ significantly from those patients who did feel fit to drive.
A comparison between ITT patients 
50 yr and older (>50 yr) showed that the younger patients had better results in number of "right answers" in COG (48.2 ± 14.5 for >50 yr vs 59.2 ± 10.6 for 50 yr, P = 0.044) and 2-Hand-Score (6.8 ± 2.6 for >50 yr vs 4.7 ± 1.2 for 50 yr, P = 0.022). Other differences could not be detected.
DISCUSSION
The use of transdermal opioid application systems is an accepted treatment for cancer and noncancer pain conditions (1–5,40). Besides transdermal fentanyl, TBD is available for these indications. However, side effects, such as impaired psychomotor function, sedation, and dizziness, have been identified as clinically relevant problems resulting from opioid treatment (11–20). Not only opioids, but pain itself, may impair psychomotor function (41). Veldhuijzen et al. (42) assessed highway driving performance in patients with nonmalignant pain and showed that patients with pain performed significantly worse than healthy volunteers. The difference in our study is that our patients were under adequate pain therapy. The patients with pain in the study of Veldhuijzen et al. (42) drove without opioid medication and the pain intensity varied from 1.2 to 9.9 cm on a 10-cm visual analog score.
Kuhajda et al. (43) demonstrated that memory performance of patients with headache is impaired. The results of our study and previous publications of our study group assessing the effects of transdermal fentanyl (13) and controlled-release oxycodone (16) on driving ability showed no impact of pain intensity on cognition and psychomotor function, when pain has been treated sufficiently.
As driving is considered to be an important aspect of a self-determined life, some publications have focused on the effect of opioids on patients' driving ability (13,16,17,21,22). Breivik (44) also placed emphasis on the importance of investigations assessing driving ability under opioid therapy for pain treatment, and gave general recommendations for doctors and patients.
Nilsen et al. (45) evaluated the effect of short-acting opioids on performance in a driving simulator. The performance of patients with chronic nonmalignant pain, with or without 150 mg codeine daily, was inferior to that of healthy volunteers. Other authors investigated the effects of long-acting opioids. A previous investigation by our study group (16) used the same study protocol as in this study and was unable to demonstrate noninferiority for the primary end-point when assessing patients treated with controlled release oxycodone compared with the control group + . In contrast to these results were the results of another investigation by our study group (13). For patients with chronic noncancer pain, with transdermal fentanyl, Sabatowski et al. (13) demonstrated noninferiority for the primary end-point. We assume that patients with transdermal opioid application systems have better test results than those receiving oral oxycodone medication due to stable plasma concentrations for the transdermal systems and lower peak concentrations, which are rarely reached (1–4). This theory is supported by other investigators. Menefee et al. (46) also tested patients with chronic nonmalignant pain under long-term therapy with transdermal fentanyl and found no impairment in driving ability, similar to the findings of Sabatowski et al. (13) and ours.
A factor that might explain the findings of Gaertner et al. (16) is the biexponential pharmacokinetic of controlled release oxycodone, with a fast release of around 30% of the drug within 1 h to provide immediate onset of pain relief. It could be argued that this rapid drug release might imitate an effect close to immediate-release opioids, a theory which is supported by the study of Kamboj et al. (47), who performed a double-blind, placebo-controlled study in a crossover design, demonstrating that patients receiving immediate-release in addition to a stable sustained-release medication had memory impairment as well as reduced performance on a complex tracking task.
However, in the buprenorphine study and in the oxycodone study, plasma levels of the opioids were not examined.
In contrast to our hypothesis are the findings of Jamison et al. (15) who compared psychomotor function of patients with noncancer pain receiving transdermal fentanyl or controlled-release oxycodone in a crossover design and reported no significant differences between the two opioids. However, the informational value of the results is limited by the fact that the authors performed only two paper tests. Byas-Smith et al. (17) evaluated the driving ability of patients with chronic nonmalignant pain under oral opioid medication and could not find a difference from healthy subjects. We have to note that the oral opioids taken by the patients were not only one substance; the 21 patients received nine different opioids. Even a patient with tramadol, a classified step two analgesic in the World Health Organization analgesic ladder, was included into the study. Furthermore, there was no benchmark defining inability to drive. As a result, the findings of these studies (15,17) are not comparable to ours.
CONCLUSIONS
The results of the present study demonstrate that the driving ability of patients with chronic noncancer pain treated with TDB was not inferior to the CG of healthy volunteers. However, because of the individual variability of test results, an individual assessment is recommended.
Footnotes
Accepted for publication July 3, 2007.
Supported by an unrestricted grant of Grünenthal GmbH, Germany.
Financial relationships between any of the authors and any commercial interest with a vested interest in the outcome of the study are disclosed.
Reprints will not be available from the author.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
