Anesth Analg 2004;98:224-229
© 2004 International Anesthesia Research Society
OBSTETRIC ANESTHESIA
The Effects of Prolonged Ambulation on Labor with Epidural Analgesia
Stéphane Frenea, MD*,
Christine Chirossel, MD*,
Raphaël Rodriguez
,
Jean-Philippe Baguet, MD, PhD
,
Claude Racinet, MD, and
Jean-Francois Payen, MD, PhD* Section Editor
*Department of Anesthesiology,
Department of Obstetrics, and
Department of Cardiology and Hypertension, Albert Michallon Hospital, Grenoble, France
Address correspondence and reprint requests to Jean-Francois Payen, Département d’Anesthésie-Réanimation, Hôpital Albert Michallon, BP 217, 38043 Grenoble, France. Address e-mail to jfpayen{at}ujf-grenoble.fr
 |
Abstract
|
|---|
Ambulation during labor is becoming more popular, although its impact on the progress of labor and on pain intensity remains unclear. We wondered whether prolonged ambulation with epidural analgesia had a possible effect on duration of labor and pain. In this prospective, randomized trial, 61 parturients with uncomplicated term pregnancies were allocated to be recumbent (n = 31) or to ambulate (n = 30). Epidural analgesia was provided with intermittent administrations of 0.08% bupivacaine-epinephrine plus 1 µg/mL of sufentanil. Of the 30 women assigned to the ambulatory group, 25 actually walked. Their ambulating time was 64 ± 34 min (mean ± SD), i.e., 29% ± 16% of the first stage. There were no differences between the two groups in the length of labor and in pain visual analog scale scores. However, the ambulatory group received smaller doses of bupivacaine (6.4 ± 2.2 mg/h versus 8.4 ± 3.6 mg/h; P = 0.01) and of oxytocin (6.0 ± 3.7 mUI/min versus 10.2 ± 8.8 mUI/min; P < 0.05). A greater ability to void was also found in the ambulatory group (P < 0.01). Although the duration of labor and pain relief was unchanged, these findings support that ambulation during labor may be advantageous.
IMPLICATIONS: This study compared the duration of labor and pain relief between parturients receiving epidural analgesia who were ambulated or were recumbent. Whereas walking had no impact on either duration of labor or pain relief, it was associated with a reduction in both bupivacaine and oxytocin requirements.
|
Introduction
|
|---|
There is general agreement that many women desire mobility during labor. However, the possible benefits of walking on labor and delivery are inconclusive. Initial reports stated that ambulation could reduce the duration of the first stage of labor, the need for labor augmentation with oxytocin, the need for analgesia, the requirement for instrumental deliveries (1–4), and the incidence of fetal distress (4). However, other studies found no benefit of standing or walking (5,6). The evaluation of walking in previous studies has been limited in duration, with ambulation being stopped when IV or epidural analgesia was required.
Because of efforts to reduce motor block associated with epidural analgesia, there is renewed interest in ambulation during labor. Although many clinical trials have compared various analgesic protocols in ambulatory women during labor (7–10), the question of the impact of ambulation on labor in parturients receiving epidural analgesia has received little attention. Two randomized trials have been conducted on nulliparous women in labor with epidural analgesia (11) or combined spinal-epidural analgesia (12). Despite a large number of included patients, these studies found that walking during labor had no detectable effect on any outcome of labor or delivery. However, in one of these studies, the time spent walking was short (<15% of the duration of the first stage of labor) (11), and in the other, <50% of women in the ambulatory group spent a significant part of the first stage of labor walking (12). In addition, sitting in a chair, instead of walking, was permitted in both studies. Therefore, all conditions might not have been fulfilled to confirm the lack of difference between ambulatory and nonambulatory groups. We hypothesized that more prolonged ambulation, i.e., more than 25% duration of the first stage of labor, may shorten the first stage of labor and reduce pain intensity. The aim of the present study was to investigate these two primary outcomes (duration of labor and pain scores) by comparing ambulation versus recumbence in women with uncomplicated term pregnancies receiving epidural analgesia.
|
Methods
|
|---|
The study protocol was approved by the Grenoble Institutional Ethical Committee. Written informed consent was obtained from ASA physical status I or II parturients who requested epidural analgesia. Nulli-parous and multiparous mothers were randomly allocated to be recumbent (Group 1) or to ambulate (Group 2) using sealed, numbered envelopes once the parturient had her first dose of epidural analgesia. Parturients were eligible for recruitment if they had a singleton pregnancy from 37 to 42 wk of gestation in a fixed cephalic, uncomplicated presentation, and 3- to 5-cm cervical dilation at the time of epidural insertion. They could be in spontaneous labor or admitted for elective induction. Patients were excluded if they had any of the following criteria: unfixed cephalic presentation, cervical dilation more than 5 cm, a contraindication to epidural analgesia, or a systolic arterial blood pressure (SAP) <100 mm Hg before epidural insertion. Other criteria for exclusion were twin pregnancy, history of cesarean delivery, and any known complication of pregnancy including breech presentation. A normal fetal heart rate pattern was also required.
Fetal heart rate was continuously monitored, and 500–1000 mL of Ringer’s lactate solution infused IV before the epidural analgesia. A lumbar epidural catheter was inserted 4–5 cm into the L3-4 or L4-5 interspace through a 18-gauge Tuohy needle (B-Braun Medical, Melsungen, Germany) with the patient in the sitting position. After a test dose of 3 mL of 2% lidocaine with 1:200,000 epinephrine, a first analgesic dose of a total of 15 mL of a solution containing 0.08% bupivacaine with 1:200,000 epinephrine and 1 µg/mL of sufentanil was injected through the epidural catheter in 2 boluses 5 min apart. The level of pain was assessed using a visual analog scale (VAS) (0 mm = no pain; 100 mm = greatest amount of pain), and adequate analgesia was defined as a VAS <30 mm. Patients were then randomly allocated to one of the two groups. In Group 1, patients were confined to bed in dorsal or lateral recumbence. In Group 2, patients were asked to walk at least 15 min of each hour or 25% duration of the first stage of labor. In the two groups, intermittent top-up of 15 mL of the 0.08% bupivacaine solution was administrated by the anesthesiologist whenever the patient had pain again. If pain persisted for 15 min after the top-up (VAS >30 mm), an additional 5 mL of 0.25% bupivacaine was given. The maximal dose of sufentanil did not exceed 30 µg. The management of analgesia was under anesthesiologist care only.
In Group 2, ambulation was permitted 15–20 min after the initial injection, provided there was no postural hypotension, no motor block in lower limbs, no proprioception impairment, and no fetal heart rate decelerations. Postural hypotension was defined as a SAP of <100 mm Hg or a decrease of more than 20% of control values at 1, 3, and 5 min after standing up (13). Motor block was defined as a detectable weakness in the legs when pushing against resistance with fully flexed knees in a supine position and then as an inability to stand up after squatting. Proprioception was assessed with distal joint proprioception (the great toe was moved up and down, and the patient was asked to state its position) and a Romberg test. The patient was accompanied by her spouse or a midwife at all times while walking around the labor and delivery suite. Fetal heart rate was continuously monitored by telemetry during ambulation (Hewlett Packard 80300A, Hewlett Packard, Paris, France). In addition, blood pressure was measured every 15 min by a noninvasive monitor (Diasys Integra, Novacor, Rueil-Malmaison, France) equipped with a movement sensor recording the posture of the patient (lying, sitting, and standing) (14). The women were asked to return to bed when they requested an epidural top-up or if they experienced weakness or sensory changes. Walking ended when examination by the midwife revealed full cervical dilation. In the recumbent group (Group 1), monitoring of labor was similar, but without telemetry. To analyze side effects in this group, arterial hypotension was defined as a SAP of <100 mm Hg or a decrease of more than 20% of control values while supine and motor block as a detectable weakness in the legs when pushing against resistance with fully flexed knees in a supine position. No Romberg test was used in this group. In the two groups, maternal blood pressure, motor power, level of pain, and cervical dilation were assessed after each top-up.
Labor was managed, according to the department’s standard policies, under midwife care. In spontaneous labor, augmentation with oxytocin was used if labor was considered ineffective. This was preceded by an amniotomy if the fetal membranes were intact. Oxytocin administration was standardized as follows: the initial dose of oxytocin was 4 mIU/min and was increased every 30 min by 4 mIU/min if required to a maximum of 40 mIU/min. The obstetrician was informed when the total amount of oxytocin exceeded 10,000 mIU. The oxytocin infusion pump was interrupted during ambulation (there was no portable infusion), which would have reduced the time available during labor to reach the larger doses of oxytocin. Therefore, a correction for the determination of the mean hourly rate of oxytocin injection (Qoxytocin) was applied to each ambulating parturient, according to the equation:
|
Delivery management was left to the discretion of the midwife or obstetrician. All patients were given 6 mL of 0.25% bupivacaine before pushing for expulsion.
Several variables were measured in the two groups to investigate the two primary outcomes (duration of labor and pain): the intervals between epidural insertion and complete cervical dilation (first stage) and to delivery (second stage). Pain VAS score was performed after each top-up; however, considering that the number of pain measurements after epidural insertion varied among subjects according to their length of labor, a mean pain VAS score was calculated during the first stage of labor. In addition, the dose of oxytocin, the hourly consumption of bupivacaine from epidural insertion to delivery (including bolus of bupivacaine for expulsion), the number of epidural top-ups, and the interval between each were determined to complete analysis of the two primary outcomes. Secondary outcomes were related to type of delivery, maternal blood pressure, ability to void, maternal side effects, and fetal status at birth (Apgar score at 1 and 5 min and umbilical arterial pH). One day after delivery, each mother was questioned regarding side effects and overall satisfaction using a 4-level scale (dissatisfied, poorly satisfied, satisfied, and extremely satisfied).
The study population size for this trial was calculated assuming a 30% reduction in the duration of first stage of labor in the ambulatory group, as reported in other studies (2,4,15). The duration of first stage of labor was 444 ± 192 min in recumbent nulliparous women (16). Based on the formula for a normal distribution and assuming a two-sided type I error of 0.05 and a power of 0.80, 26 patients in each of the 2 groups were required. To allow for parturients who might not complete the study, we decided to enroll a minimum of 60 parturients. Data were expressed as mean ± SD, except for pain VAS scores (median; 95% confidence interval). Frequency data were compared using the 
2 test (StatView SE program, SAS Institute Inc, Cary, NC). Discrete data were compared using the Student t-test. Pain VAS scores were compared using the Mann-Whitney U-test. Evolution of arterial blood pressure was assessed using one-way analysis of variance (ANOVA) for repeated measurements. Each value was compared with that obtained at the control period using the Scheffé F-test as a post hoc test. Statistical significance was declared when P < 0.05.
|
Results
|
|---|
Between February 1998 and March 1999, 62 parturients were enrolled in this study. Of these, one patient was excluded because of incomplete data (from Group 1). Therefore, 61 parturients completed the study: 31 Group 1 and 30 in the Group 2. Of Group 2, two patients did not walk because a cesarean delivery was required before walking and three because of postural hypotension despite treatment with IV ephedrine and a fluid bolus. The mean walking time of the 25 parturients who actually walked was 64 ± 34 min (mean ± SD). This time corresponds to 29% ± 16% of the interval from epidural insertion to complete cervical dilation (first stage). No significant difference in this relative walking time was found between nulliparous and multiparous women.
The results were analyzed on an intention-to-treat basis. The characteristics of the women in the two groups were similar with respect to age, weight, height, gestation, parity, percent of elective induction, and cervical dilation at the time of epidural insertion (Table 1). Table 2 shows the labor and delivery outcomes in the two groups. Most of the characteristics of labor and delivery were not significantly different between the two groups, including the length of labor, the cervical dilation rate, the need for oxytocin, the mode of delivery, and the infant’s outcomes. The overall length of labor, including the eight patients (n = 3 in Group 1 and n = 5 in Group 2) who did not reach complete cervical dilation because of cesarean delivery, did not differ between the two groups. There was no abnormality of fetal heart rate during ambulation. Three babies were born with an Apgar score <7 at 1 min (n = 2 in Group 1 and n = 1 in Group 2) that was resolved at 5 min. However, the consumption of bupivacaine and the doses of oxytocin were significantly smaller in Group 1 than in Group 2 (Table 2). The reduction in analgesic requirements in Group 2 was not accompanied by a difference in pain VAS scores between the two groups (Fig. 1). The number of additional boluses of bupivacaine 0.25% was similar in the 2 groups (0.6 ± 1.3 versus 0.4 ± 0.7, respectively).
For the 51 parturients who had oxytocin augmentation, the reduction in oxytocin consumption in Group 2 was not accompanied by a difference in the overall length of labor between groups (309 ± 164 versus 318 ± 131 min; P = 0.83). It should be noted that the 51 patients with oxytocin augmentation had a longer duration of labor (313 ± 148 min versus 212 ± 140 min; P = 0.05) and a slower cervical dilation rate (1.9 ± 1.1 cm/h versus 3.6 ± 2.3 cm/h; P < 0.01) than patients with no oxytocin augmentation (n = 10); no difference in analgesic requirement was found between these two populations. In the five parturients in the ambulatory group that did not actually walk (because of postural hypotension or cesarean delivery), oxytocin and bupivacaine use (6.9 ± 4.0 mIU/min and 7.2 ± 2.2 mg/h, respectively) was comparable with that of the parturients who walked.
Table 3 indicates maternal side effects. The number of hypotensive episodes did not differ between groups. At no stage of labor were SAP and diastolic blood pressures statistically different between groups. Two patients in the ambulatory group had an initial motor block, but it resolved quickly, and they subsequently walked. Proprioception impairment was limited to one parturient per group. No difference in other side effects was found. Conversely, ambulation led to a significantly larger number of patients able to spontaneously void and to less requirements for bladder catheterization in the last hour (Table 3).
The interview on the day after delivery revealed that many women were extremely satisfied in both groups (19 versus 22; P = 0.56) and that most of the ambulatory patients (28 of 30) would choose to walk again during a future labor. The numbers of subjects with headache (n = 5), backache (n = 12), drowsiness (n = 7), or urinary retention requiring catheterization (n = 2) were not different between the two groups. These headaches resolved spontaneously.
|
Discussion
|
|---|
The present study indicates that ambulation for more than 25% of first stage of labor under epidural analgesia did not significantly change the length of the labor or other outcomes of labor and delivery. This is in agreement with several other studies performed on ambulation (6,11,12,17). However, the lack of effect on both primary outcomes (duration of labor and pain scores) is balanced by smaller doses of bupivacaine and of oxytocin in ambulating women. These unexpected results may revive interest in walking during epidural analgesia for uncomplicated term pregnancies.
We hypothesized that prolonged ambulation with analgesia could shorten the first stage of labor. This issue remained inconclusive from previous studies because fewer than half of the patients who had been assigned to walk did so (11,12). In addition, the time spent walking was short (25 ± 23 min) (11) or defined as a minimum of five minutes walking per hour (17). Therefore, we encouraged women to walk as much as possible, under continuous monitoring of the fetal heart rate and with an attendant at all times while ambulating. Most of our parturients (25 of 30) walked for 29% ± 16% of the duration of their first stage of labor, which was somewhat less than our initial goal. Despite this, no significant change in the length of labor was found between the two groups. From these results and those of randomized trials without epidural analgesia (5,6), one can conclude that the length of labor is too coarse a criterion to detect any difference between postures and that it depends on numerous factors in addition to posture.
Among these other factors, the use of oxytocin is critical because there is a relationship between the dose of oxytocin and the length of labor (18). A longer duration of labor was found in all of the 51 women receiving oxytocin augmentation, indicating that oxytocin was required for ineffective labor regardless of the group. Although the number of women requiring oxytocin for labor augmentation was similar between the two groups, our ambulatory group received smaller total doses of oxytocin than did the recumbent group. Most studies did not take into account the hourly dose of oxytocin (6,11,12,17). In this study, doses of oxytocin were calculated to complete the analysis of the primary outcome, i.e., duration of labor. Although they were responsible for the oxytocin management, the midwives did not receive specific instruction regarding the ambulatory group. As a result, the overall length of labor was similar between groups. In addition, oxytocin augmentation did not affect the requirement for bupivacaine. Our results suggest that the lack of benefit on duration of labor may be offset by the use of smaller doses of oxytocin in the ambulatory group. In a small, randomized trial, similar effects on labor enhancement were found among eight ambulatory patients and six patients with oxytocin infusion (3). However, the mechanism is not clear because the effect of posture on uterine contractility remains controversial (1,15).
Surprisingly, the ambulatory group received less bupivacaine than the recumbent one, whereas pain VAS scores did not differ between the two groups. Top-ups were given by the anesthesiologist whenever the patient complained that pain had recurred. Such a method can be as effective as patient-controlled epidural analgesia to give pain relief and ability to walk (7). As a result, no discomfort was found in either group. This indicates that ambulation could have a sparing effect on the analgesic requirement during labor. Conversely, others did not find such a sparing effect of walking (6,11,12). There are some issues that may explain this discrepancy: the absence of pain measurement (6,12), the request for analgesics only when patients are no longer ambulated (6), the use of continuous epidural infusion in ambulatory and nonambulatory groups (11), or the uncontrolled choice of posture in the ambulatory group (walking, standing, or sitting) as well as in the control group (sitting or lying) (12). However, less pain with an upright than with a supine position in parturients before epidural analgesia has been reported (19,20). At present, one can only speculate about the possible analgesic effect of ambulation. Standing or ambulating could affect the pain perception by a possible reduction in the traction and pressure on roots of the lumbosacral plexus, on the adnexa, or on skeletal muscle in the pelvis during uterine contractions (20). Further studies are required to answer these questions.
No negative aspects of allowing parturients to ambulate were found with respect to the incidence of arterial hypotension, motor block, and proprioception impairment in agreement with other studies (6–10). However, persistent postural hypotension was found in three parturients in Group 2, which precluded their further walking. There was no evidence of bupivacaine overdosing in these three women because no motor block, no reduced consumption rate of bupivacaine, or lengthening of interval between each top-up was found. This proportion of patients with postural hypotension was comparable with other reports (21,22), and no adverse effect was noted on the fetal heart rate. However, this indicates that blood pressure needs to be carefully checked after each epidural top-up before allowing ambulation.
Few parturients (6 of 61) had motor block, and this did not last long enough to preclude subsequent walking in the ambulatory group. This positive aspect was primarily because of the use of a small concentration of bupivacaine. Several authors have recommended the use of opioids (fentanyl or sufentanil) either alone or in association with dilute local anesthetic solution (0.08%–0.125% bupivacaine) to allow ambulation after epidural or combined spinal-epidural analgesia (8–10,23). In addition, intermittent top-ups reduce the total use of bupivacaine and consequently, the incidence of motor block, compared with continuous infusion (7). However, our incidence of motor block might have been reduced further if we had used dilute ropivacaine (0.08%) (9) or no test dose (10).
Finally, the greater ability to void could be another benefit of ambulation during labor, along with the feelings of freedom and self-control. It is well known that lying in a recumbent position makes voiding on a bedpan more difficult, as verified in this study. Previously, it has been reported that more than one third of ambulatory patients could void while receiving combined spinal-epidural analgesia (13). In addition, all women receiving epidural analgesia with 0.08% ropivacaine were able to void (9). In our study, ability to void was not caused by the reduced bupivacaine consumption in Group 2 because it has been shown that there was no difference in voiding between ambulatory groups receiving either 0.0625% or 0.125% bupivacaine (10). Moreover, because of the small incidence of motor block in the recumbent group, the incidence of spontaneous micturition could have been more frequent in this group if women had been asked to do this. However, the advantage of spontaneous micturition was not reflected by a dramatic reduction in bladder catheterization in Group 2. This may be because of incomplete voiding during labor or to the midwife’s habits. Nevertheless, ambulatory women were very grateful to be able to void in the bathroom.
In conclusion, this study showed that ambulation with epidural analgesia did not change the duration of labor and pain scores but was associated with reduced doses of oxytocin and analgesics. Ambulation could be proposed to many parturients with uncomplicated term pregnancies, providing that strict criteria are met and constant assistance is provided. Finally, having the choice to ambulate is well accepted by women, even if they finally prefer to remain in bed once pain is relieved by epidural analgesia.
|
Acknowledgments
|
|---|
The authors thank Jonathan Coles for helpful comments on the manuscript.
|
References
|
|---|
- Mendez-Bauer C, Arroyo J, Garcia Ramos C, et al. Effects of standing position on spontaneous uterine contractility and other aspects of labor. J Perinat Med 1975; 3: 89–100.[Medline]
- Stewart P, Calder AA. Posture in labour: patients’ choice and its effect on performance. Br J Obstet Gynaecol 1984; 91: 1091–5.[ISI][Medline]
- Read JA, Miller FC, Paul RH. Randomized trial of ambulation versus oxytocin for labor enhancement: a preliminary report. Am J Obstet Gynecol 1981; 139: 669–72.[ISI][Medline]
- Flynn AM, Kelly J, Hollins G, Lynch PF. Ambulation in labour. BMJ 1978; 2: 591–3.
- McManus TJ, Calder AA. Upright posture and the efficiency of labour. Lancet 1978; 1: 72–4.[ISI][Medline]
- Bloom SL, McIntire DD, Kelly MA, et al. Lack of effect of walking on labor and delivery. N Engl J Med 1998; 339: 76–9.[Abstract/Free Full Text]
- Collis RE, Plaat FS, Morgan BM. Comparison of midwife top-ups, continuous infusion and patient-controlled epidural analgesia for maintaining mobility after a low-dose combined spinal-epidural. Br J Anaesth 1999; 82: 233–6.[Abstract/Free Full Text]
- Connelly NR, Parker RK, Vallurupalli V, et al. Comparison of epidural fentanyl versus epidural sufentanil for analgesia in ambulatory patients in early labor. Anesth Analg 2000; 91: 374–8.[Abstract/Free Full Text]
- Campbell DC, Zwack RM, Crone LA, Yip RW. Ambulatory labor epidural analgesia: bupivacaine versus ropivacaine. Anesth Analg 2000; 90: 1384–9.[Abstract/Free Full Text]
- Cohen SE, Yeh JY, Riley ET, Vogel TM. Walking with labor epidural analgesia: the impact of bupivacaine concentration and a lidocaine-epinephrine test dose. Anesthesiology 2000; 92: 387–92.[ISI][Medline]
- Vallejo MC, Firestone LL, Mandell GL, et al. Effect of epidural analgesia with ambulation on labor duration. Anesthesiology 2001; 95: 857–61.[ISI][Medline]
- Collis RE, Harding SA, Morgan BM. Effect of maternal ambulation on labour with low-dose combined spinal- epidural analgesia. Anaesthesia 1999; 54: 535–9.[ISI][Medline]
- Breen TW, Shapiro T, Glass B, et al. Epidural anesthesia for labor in an ambulatory patient. Anesth Analg 1993; 77: 919–24.[Abstract/Free Full Text]
- Mallion JM, Mouret S, Baguet JP, et al. Ambulatory blood pressure variation in normotensive subjects in relation to the sitting or standing position. Blood Press Monit 2000; 5: 169–73.[ISI][Medline]
- Chen SZ, Aisaka K, Mori H, Kigawa T. Effects of sitting position on uterine activity during labor. Obstet Gynecol 1987; 69: 67–73.[Abstract/Free Full Text]
- Williams RM, Thom MH, Studd JW. A study of the benefits and acceptability of ambulation in spontaneous labour. Br J Obstet Gynaecol 1980; 87: 122–6.[ISI][Medline]
- Nageotte MP, Larson D, Rumney PJ, et al. Epidural analgesia compared with combined spinal-epidural analgesia during labor in nulliparous women. N Engl J Med 1997; 337: 1715–9.[Abstract/Free Full Text]
- Frigoletto FD Jr, Lieberman E, Lang JM, et al. A clinical trial of active management of labor. N Engl J Med 1995; 333: 745–50.[Abstract/Free Full Text]
- de Jong PR, Johanson RB, Baxen P, et al. Randomised trial comparing the upright and supine positions for the second stage of labour. Br J Obstet Gynaecol 1997; 104: 567–71.[ISI][Medline]
- Melzack R, Belanger E, Lacroix R. Labor pain: effect of maternal position on front and back pain. J Pain Symptom Manage 1991; 6: 476–80.[ISI][Medline]
- Collis RE, Baxandall ML, Srikantharajah ID, et al. Combined spinal epidural analgesia with ability to walk throughout labour. Lancet 1993; 341: 767–8.
- Shennan A, Cooke V, Lloyd-Jones F, et al. Blood pressure changes during labour and whilst ambulating with combined spinal epidural analgesia. Br J Obstet Gynaecol 1995; 102: 192–7.[ISI][Medline]
- Sia AT, Chong JL, Tay DH, et al. Intrathecal sufentanil as the sole agent in combined spinal-epidural analgesia for the ambulatory parturient. Can J Anaesth 1998; 45: 620–5.[Abstract/Free Full Text]
Accepted for publication July 25, 2003.
Top
Abstract
Introduction
, Group 1; 
, Group 2.
