| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
a Section of Orthopaedic Surgery and Rehabilitation Medicine, Department of Surgery, The University of Chicago, Illinois
b Department of Anesthesiology and Critical Care, The University of Chicago, Illinois
c Department of Health Studies, The University of Chicago, Illinois
Louis F. Draganich, The University of Chicago, Section of Orthopaedic Surgery and Rehabilitation Medicine, Department of Surgery, MC 3079, 5841 South Maryland Avenue, Chicago, IL 60637 E-mail: ldragani{at}surgery.bsd.uchicago.edu.
Decision Editor: John E. Morley, MB, BCh
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
Methods. A randomized, crossover, four-period, double-blind, placebo-controlled laboratory trial was performed. Twelve healthy elderly subjects (average age, 67 years; range, 65–72 years) were tested. Subjects were assigned the three antidepressant drugs or a placebo in a random order. Single doses of amitriptyline 50 mg, desipramine 50 mg, paroxetine 20 mg, or placebo were given 4 hours prior to gait testing. Temporal-distance measures and kinematics of the lower trailing limb (i.e., limb going over obstacle last) were obtained.
Results. Compared with placebo, amitriptyline significantly reduced gait velocity by as much as 8.0% (p = .028), cadence by as much as 4.9% (p = .012), angular velocity of hip flexion by as much as 10.0% (p = .004), and angular velocity of knee flexion by as much as 8.3% (p = .018) during the crossing strides when stepping over obstacles. Except for knee flexion angle, unobstructed gait was not affected. Amitriptyline affected integrative capacity of the central nervous system (CNS) and ability to concentrate as measured by psychomotor and mood tests.
Conclusions. The results for amitriptyline suggest that the subjects slowed their obstacle crossing speeds as a result of reduced CNS integrative capacities. Neither paroxetine nor desipramine significantly affected gait, psychomotor function, or mood.
FALLS are the leading cause of death from injury in people 65 years or more of age (1). Most falls by elderly persons occur during activities such as walking or changing position, and tripping or tripping over an obstacle are among the most reported causes of falls (2)(3)(4). Antidepressants have been found to be associated with falling (5). One of the possible reasons for increased risk of falling with antidepressants may be drug-induced gait abnormalities, which formed the rationale for performing the present study.
Many studies have documented acute and subacute impairment in psychomotor performance with the tricyclic antidepressant (TCA) amitriptyline (6). Other TCAs have been shown to have no effects or only mildly impairing effects on psychomotor or cognitive performance (7). There is a developing body of literature that indicates that a newer generation of antidepressants, the selective serotonin reuptake inhibitors (SSRIs), are not associated with psychomotor or cognitive impairment or subjective sensations of drowsiness (8), and are as efficacious as the TCAs in alleviating depression (9). Thus, SSRIs have been predominantly used in recent years to treat depression. However, recent studies have found SSRIs to be just as or more likely than the TCAs to be associated with falls (10)(11).
We hypothesized that single doses of amitriptyline (a tertiary-amine TCA), desipramine (a secondary-amine TCA), and paroxetine (an SSRI) would affect the kinematics of the trailing limb when stepping over obstacles. We also hypothesized that amitriptyline would affect mood and psychomotor performance and that desipramine and paroxetine would not.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
The University of Chicago Institutional Review Board approved the protocol, and written informed consent was obtained. In order to reduce any expectations the subjects may have had about the drug they might receive, the consent form included several drugs or classes of drugs that subjects were told could be administered in the study (e.g., antiarrhythmic agents, stimulants, anxiolytics/sedatives, antidepressants, antihistamines, or placebo). Subjects were asked to agree not to take any other drugs during their participation in this study.
Study Design
Subjects arrived at the General Clinical Research Center (GCRC) by livery service at approximately 6:15 AM. At 6:30 subjects were given a baseline resting EKG and then were given practice on the psychomotor tests. At 7:15 AM subjects were given a standardized breakfast adjusted for body weight. At 7:45 AM baseline testing of mood and psychomotor performance commenced. Drugs were administered with 100 cc to 200 cc of water at 8:00 AM. Doses of paroxetine 20 mg, amitriptyline 50 mg, desipramine 50 mg, or placebo were tested in a four-period, double-blind, crossover design. Treatment orders were randomly assigned from those residing in three 4 x 4 Latin squares. The washout periods between treatments were at least 6 days. Drug and placebo capsules were identical in size, weight, and appearance. The randomization schedule, indexed by subject number and trial number, was generated by the project statistician at the beginning of the study and given to the pharmacist. The pharmacist placed the appropriate capsule in a white envelope with the subject number and trial number noted on it and gave it to the nurse in the GCRC on the morning that it was to be administered. The treatment team was blind to treatment assignment.
Gait testing began at noon, 4 hours after the drug had been ingested and when drug plasma levels were predicted to be at or close to peak levels. The gait testing took approximately 45 minutes to perform (see below). The subjects performed several practice trials of unobstructed gait along the walkway before the formal tests of gait began. The technician conducting the sessions did not provide any feedback or discuss any aspect of the study other than providing instructions necessary to perform the task.
Psychomotor tests were given 15 minutes before ingestion of the drug and at 1, 3, 5, 7, and 12 hours after ingestion. Mood tests were given 15 minutes before ingestion of the drug and every hour after ingestion with the last test given 12 hours after ingestion. The subjects remained in the hospital overnight. They were given a resting EKG and were evaluated by a physician before being taken home by livery service the morning after the day of the test session.
Testing Series
Gait..--
The motion analysis laboratory included a 9.5-m walkway. The kinematic parameters were measured using the Watsmart (Northern Digital, Inc., Waterloo, Ontario, Canada) optical electronic three-dimensional digitizing and analysis system, which detected the positions of infrared light-emitting diodes. Rigid segments of infrared light-emitting diodes were attached using elastic straps to the foot, shank, thigh, and pelvis of the left lower extremities of the subjects. The motion data were acquired at a rate of 100 samples per second. The temporal-distance measurements of gait consisted of stride length normalized to lower limb length, walking velocity, and cadence. This methodology has been reported previously from our laboratory (15)(16).
Gait analysis was performed on all subjects for all four treatments for both unobstructed and obstructed gait. The subjects were dressed in shorts and their own comfortable sneakers or shoes with low heels. Anthropometric measurements were obtained.
Three sets of experiments were performed in order to challenge the subject in increasing degrees of difficulty. Subjects first performed several practice trials. In experiment 1, the subject was asked to walk on the walkway, without encountering obstacles, in a self-selected manner. Each subject's average step length over three to six steps was determined. In experiments 2 and 3 the subject stepped over an obstacle (Fig. 1). Before performing experiments 2 and 3 we determined the subject's final starting position and baseline location for the obstacle that allowed the subject to consistently step within 5 cm of a target with the trailing foot, step over the obstacle first with the leading foot, then with the trailing foot, and continue walking along the walkway, all in a self-selected manner. The obstacle spanned the walkway and consisted of a white plastic dowel, 6 mm in diameter, placed across an aluminum frame at heights of 102, 153, or 204 mm (15)(16). In experiment 2, the subject was asked to begin walking comfortably from the final starting position, step over the obstacle (placed at the baseline location) for each of three heights in a self-selected manner, and continue walking to the end of the walkway, all in a smooth and continuous fashion. Trials for the three heights were performed in a random order. In experiment 3, the obstacle of each height was selected and placed at a distance of 10%, 20%, or 30% of the subject's step length, measured in experiment 1. Trials for the nine height and distance combinations were performed in a random order. The subject was instructed to begin walking from the final starting position, maintain the same stride as in experiment 2 until the target was reached, step on the target with the great toe of the left foot, step over the obstacle, and walk to the end of the walkway.
|
Data from the middle stride from each of three trials of unobstructed gait were averaged. For obstructed gait, only trials were accepted for which the toe of the subject's trailing foot landed within 5 cm of the target, defined as a successful foot placement. Three successful trials were collected for each task (i.e., for each height and for each height at each distance). Between 45 and 50 trials were performed, including those for successful obstructed and unobstructed gait trials, repeated trials, and practice trials. Data from the crossing stride from each of three successful trials of obstructed gait were averaged. The gait data obtained for stepping over the obstacle in a self-selected manner were averaged over all three heights. The data obtained for stepping over the obstacle positioned at various heights and distances were averaged over all nine heights and distances.
A safety harness, attached to a trolley that was attached to a concrete ceiling, was secured to each subject to catch the subject in case of a fall. Subjects were asked to sit and rest for 1 to 2 minutes every 10 minutes to prevent fatigue.
Mood..-- A locally developed visual analog scale (VAS) was used to measure the effects of the treatments on mood. The VAS consisted of twenty-three 100-mm lines, each labeled with an adjective (e.g., clumsy, difficulty concentrating, dizzy, dry mouth, feel drug effect, sleepy). Subjects were instructed to place a mark on each line indicating how they felt at the moment, ranging from "not at all" to "extremely."
Psychomotor..-- Five sets of psychomotor tests were performed: Digit Symbol Substitution Test (17), Backward Digit Span Test (17), eye–hand coordination test (18), auditory reaction time test (18), and Critical Flicker Fusion Test (CFFT). Only the last test will be briefly described as the other tests failed to detect any drug effects. The CFFT requires judgment as to whether a light stimulus that is flickering at a constantly changing rate is flickering or steady (fusion). The test assesses overall integrative capacity of the CNS and more specifically is a measure of alertness (19). A decreased threshold (i.e., decreased ability to perceive a flickering light) indicated impairment. Three ascending (flicker-to-fusion) and three descending (fusion-to-flicker) series were performed.
Statistical Treatment
All variables were separately analyzed by repeated measures analysis of variance (ANOVA); p levels 
.05 were regarded as statistically significant. The Greenhouse-Geiser adjustment to the degrees of freedom was used to assess significance levels on the basis of the overall F test. If the overall F test was significant, pairwise comparisons among treatment means were performed using paired t tests for the variables of gait, and Tukey tests were performed for the variables of psychomotor function and mood.
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Gait
Unobstructed..--
None of the subjects tripped during unobstructed gait. Compared with placebo, amitriptyline reduced maximum knee flexion during the swing phase of gait by 4.2%. This was marginally significant (p = .051). None of the drugs were found to have other statistically significant effects on unobstructed gait (Table 1 ).
|
|
Compared with placebo, amitriptyline significantly affected three dependent variables during the crossing stride, reducing cadence by 4.1%, angular velocity of hip flexion by 9.3%, and angular velocity of knee flexion by 8.3% (Table 3 ). Amitriptyline was found to differ significantly, or marginally significantly, compared with paroxetine and desipramine, on these dependent variables.
|
Psychomotor
Compared with placebo, amitriptyline was found to significantly affect the subjects' psychomotor performance on the descending series of the CFFT (p = .0001). Post hoc testing revealed that 3 hours after amitriptyline ingestion, subjects' fusion-to-flicker threshold was significantly lower relative to placebo ingestion, and that this impairment lasted for the remainder of the session. None of the other psychomotor tests was affected by any of the drugs.
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
The decreases found in angular velocity of knee flexion and angular velocity of hip flexion for amitriptyline when stepping over the obstacles were consistent with the reduced velocity of the body during the crossing stride. These decreases in angular velocity of hip and knee flexion are important for the following reasons. During gait, the hip of the lower extremity crossing over the obstacle flexes, moving the swing foot anteriorly and elevating it. At the same time, the knee of the lower extremity crossing over the obstacle also flexes, contributing to the elevation of the swing foot. Thus, as the body moves closer to the obstacle, the angular velocities of hip and knee flexion must be fast enough in order for the swing foot to cross over the obstacle without contacting it. The decreases in angular velocity of the hip and knee might be expected to decrease the vertical height of the toe of the trailing foot when it is directly over an obstacle. Compared with placebo, amitriptyline indeed decreased the height of the toe of the swing foot when the toe was directly over the obstacle by 9.9 mm when stepping over the obstacles in a self-selected manner and by 8.5 mm when stepping over the obstacles of various heights and locations. However, these decreases in toe-obstacle clearance were not significant because of the variability in the data.
The results supported our hypotheses that amitriptyline would affect mood and psychomotor performance and that desipramine and paroxetine would not. Our psychomotor results are consistent with previous reports that have found amitriptyline to affect psychomotor function (21) and desipramine and paroxetine to have little or no effect (22)(23).
Studies have found the SSRIs to be just as or more likely to be associated with falls as the TCAs (10)(11). The lack of any adverse effects of paroxetine on gait performance in the present study supports the concern raised by others (5)(24) that there may have been bias in those studies (10)(11) toward preferentially prescribing SSRIs, which were thought to be safer, to the high-risk patients (i.e., to those who had fallen most often prior to taking SSRIs). In that case, the high-risk patients may have continued to have a high fall rate, irrespective of the treatment with SSRIs. However, the lack of any adverse effects of paroxetine on gait performance in the present study may also have been due to the single dose of paroxetine tested.
In conclusion, the results of this study demonstrate causal effects for amitriptyline on obstructed gait, producing abnormal gait. Thus, the use of amitriptyline in elderly patients should be avoided if possible. However, we found no deleterious effects on gait with desipramine or paroxetine, suggesting that desipramine and paroxetine are safer to use in elderly patients with regard to gait.
|
Acknowledgments |
|---|
Received February 21, 2000
Accepted February 23, 2000
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. E. Morley Editorial: Drugs, Aging, and the Future J. Gerontol. A Biol. Sci. Med. Sci., January 1, 2002; 57(1): M2 - 6. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
