A Prospective Randomized Comparison of Epidural Infusion of Fentanyl and Intravenous Administration of Morphine by Patient-Controlled Analgesia After Radical Retropubic Prostatectomy
PATRICK H. ALLAIRE, M.D., JOSEPH M. MESSICK, Jr., M.D., Department ofAnesthesiology; JOSEPH E. OESTERLING, M.D., Department of Urology; DAVID E. BYER, M.D., Department ofAnesthesiology; ROBERT P. MYERS, M.D., MICHAEL M. LIEBER, M.D., Department of Urology; ROBERT C. CHANTIGIAN, M.D., JEFFREY O. WELNA, D.O., Department ofAnesthesiology; DAVID E. PATTERSON, M.D., MICHAEL L. BLUTE, M.D., Department of Urology; KENNETH P. OFFORD, M.S., JENNIFER A. FERGUSON, B.S., Section ofBiostatistics
In a prospective, randomized study, continuous infusion of epidural fentanyl citrate (group E) was compared with patient-controlled intravenously administered morphine sulfate (group P) for analgesia in 66 men after radical retropubic prostatectomy. Although both methods provided satisfactory analgesia, the mean comfort level scores were lower (that is, greater comfort) in group E than in group P at all observation times. The difference in mean resting comfort level scores between groups E and P was statistically significant (P~O.05) at 9 of the 11 observation times. In addition, significant differences in comfort level scores were noted at 8 of the 11 observation times during deep breathing,S of 11 during coughing, and 3 of 9 during ambulation. Maximal and minimal comfort level scores recorded by each patient during the course of the study were significantly lower (that is, less pain) in group E than in group P for all four categories of activity. The percentage of patients who reported no pain was significantly higher in group E than in group P at 9 of 11 observation times during resting and 5 of 11 observation times during deep breathing. No significant differences were noted in side effect profiles or duration of hospital stay. In summary, when two effective methods of analgesia used after radical retropubic prostatectomy were compared prospectively, patients who received epidural infusion of fentanyl were more comfortable than those with patient-controlled intravenous administration of morphine, as evidenced by lower mean, maximal, and minimal comfort level scores and a greater proportion of patients with complete relief of pain.
Postoperative pain is unpleasant and associated with potentially detrimental physiologic effects. The intermittence of the analgesia associated with intramuscular or subcutaneous injections of narcotics'> has been attributed to fluctuating and unpredictable blood concentrations of these agents. The series of steps and attendant delays between the
Address reprint requests to Dr. J. M. Messick, Jr., Department of Anesthesiology, Mayo Clinic, Rochester, MN 55905. • Mayo Clin Proc 67:1031-1041,1992
patient's request for analgesia and the actual achievement of analgesia have been reviewed. 1 Recent advances in the management of postoperative pain have included on-demand patient-controlled analgesia (PCA)-usually accomplished by intravenous administration of a narcotic-and epidural infusion of a narcotic, a local anesthetic agent, or a combination of both.t' Perceived greater effectiveness of both intravenous peA and epidural infusion of narcotics for management of postoperative pain has led to their increased use for analgesia after various surgical procedures, including
1031
1032
ANALGESIA AFTER RADICAL RETROPUBIC PROSTATECTOMY
radical retropubic prostatectomy. Potential side effects of narcotics, including respiratory depression, may occur with any route of administration. A minimal risk of mechanical failure of the pump device with either PCA or epidural infusions and the need for a specially trained physician to accomplish epidural analgesia are other factors to be considered. A recent retrospective review of 76 patients who underwent radical retropubic prostatectomy indicated a significant decrease in duration of hospital stay postoperatively among patients who received a narcotic agent epidurally in comparison with those who received narcotics intramuscularly or intravenously (PCA).6 Prospective studies in which intravenously administered opioid PCA was compared with epidural infusion of an opioid in the treatment of pain after either an orthopedic surgical procedure or cesarean section have been reported.>'? We are unaware of any prospective randomized studies that have compared continuous epidural infusion of fentanyl citrate with intravenous administration of morphine sulfate by PCA in the treatment of pain after radical retropubic prostatectomy. The current investigation was designed to compare these two modalities of analgesia after this procedure at our institution in a randomized prospective study.
MATERIAL AND METHODS This study was approved by the Mayo Institutional Review Board. Seventy-one men scheduled to undergo bilateral pelvic lymph node dissection and radical retropubic prostatectomy for clinically localized prostatic cancer, who were judged by their anesthesiologists to be eligible to receive either morphine administered intravenously (PCA) or epidural infusion of fentanyl for postoperative analgesia, were randomized prospectively to receive one of these analgesic regimens. The surgical procedure performed on all patients was the same. Patients received either general anesthesia (thiopental sodium, isoflurane in a nitrous oxide-oxygen mixture, and intravenously administered morphine or fentanyl) or epidural anesthesia (lidocaine plus intravenous supplementation with fentanyl or midazolam hydrochloride [or both] as indicated). Epidural catheters were placed in the lumbar area (L2-3 or L3-4 interspace) by the staff anesthesiologist either preoperatively for intraoperative use or in the recovery room if the patient had received a general anesthetic agent and was to receive an epidural infusion of fentanyl postoperatively. Randomization proceeded in the following sequence. The anesthesiologist determined the patient's eligibility for participation, discussed the study, obtained informed consent, and decided in conjunction with the patient on either general or regional (epidural) anesthesia for the procedure. These two anesthesia groups then were considered as strata.
Mayo Clio Proc, November 1992, Vol 67
Within each stratum, patients were randomized to receive narcotics postoperatively by either epidural infusion or PCA such that after every six patients each group had three. After induction and stabilization of anesthesia, the anesthesiologist obtained the patient's random postoperative analgesia assignment. For discussion purposes, we defined four groups by a combination of strata and postoperative analgesia assignment. Group I patients (N =24) received general anesthesia intraoperatively and randomly assigned epidural infusion of fentanyl for postoperative analgesia. Group 2 patients (N = 11) received epidural anesthesia intraoperatively and randomly assigned epidural infusion of fentanyl for postoperative pain. Group 3 patients (N = 2S) received general anesthesia intraoperatively and randomly assigned intravenously administered morphine (hereafter referred to as PCA-morphine) for postoperative relief of pain. Group 4 patients (N = 11) received epidural anesthesia intraoperatively and randomly assigned PCA-morphine for postoperative analgesia; in these patients, the epidural catheters were removed in the recovery room. During statistical analyses, it became evident that choice of general or epidural anesthesia had no substantial effect on the efficacy of postoperative analgesia. Therefore, for further analysis, groups 1 and 2 were combined into group E (epidural), and groups 3 and 4 were combined into group P (PCA-morphine). For epidural analgesia, fentanyl was administered as a bolus (usually 100 ug), followed by a 1 ug/kg per h continuous infusion of fentanyl (10 ug/ml) in the recovery room. Postoperative management of these patients included nursing care based on guidelines developed for patients who receive epidural analgesia and monitoring by pulse oximetry for 12 hours after the patient returned to the hospital room. epidural infusions of fentanyl were managed by the Acute Pain Service in the Department of Anesthesiology. Patients in the PCA-morphine group received incremental 2- to 4-mg doses of morphine administered intravenously until comfortable in the recovery room, followed by intravenous infusion of morphine by a programmable pump in the hospital room. PCA orders were written by the primary surgical service and were not influenced or controlled by the Acute Pain Service. Typical orders were for a 2-mg bolus dose of morphine with a 10- to IS-minute lockout interval and a maximal dosage of 20 to 30 mg during a 4-hour period. Comfort levels of the patients, drug side effects and related observations, duration of hospital stay, and total dose of narcotics were compared between groups. Comfort Level of Patients.-Three visual analog scales were used to evaluate the patient's comfort. Each scale consisted of a lO-cm horizontal line on a 3- by S-inch (7.6by 12.7-cm) card. On one scale, patients were asked to evaluate their level of comfort at rest by placing a vertical
.11
Mayo Clin Proc, November 1992, Vol 67
ANALGESIA AFTER RADICAL RETROPUBIC PROSTATECTOMY
mark on an unmarked line, using a scale with no pain (0) at one end and the worst pain imaginable (10) at the other. On a second scale, patients were asked to represent the quality of reliefofpain by placing a vertical mark on an unmarked line designated "complete pain relief' (0) at one end and "no pain relief' (10) at the other. A third time-referenced scale asked patients to represent their comfort level relative to the first night after the operation by placing a vertical mark on a -5 to +5 scale, on which -5 ="much worse," 0 ="no change," and +5 = "much better." In addition, patients were asked to provide verbal ratings on a 0 to 10 scale (0 = no pain and 10 = worst pain imaginable) for their level of comfort when they took a deep breath, coughed, and ambulated. Patients were not asked to evaluate their pain while sitting because some surgeons preferred that their patients not sit in a straight chair for the first few days postoperatively. The presence or absence of pruritus, nausea, vomiting, and bladder spasm was scored on a 0 to 2 scale (0 = none present, 1 = present but not necessitating medication, and 2 = present and medication needed). Patients were assessed while awake at 1600 and 2000 hours on the day of the surgical procedure and at 0800, 1200, 1600, and 2000 hours daily until 0800 hours on the third postoperative day. In accordance with previous studies of postoperative analgesia and to allow our patients to sleep during the night, we did not evaluate pain scores after 2000 hours or before 0800 hours." Physical Features and RespiratoryData.-Age, weight, height, and physical status of the patients were recorded (Table I). In all patients, a preoperative arterial blood specimen was obtained for gas analysis during breathing of room air before induction of anesthesia in the operating suite. Oxygen saturation, arterial oxygen tension, arterial carbon dioxide tension, pH, base excess, sodium, hemoglobin, and creatinine measurements from the preoperative blood sample were compared with results from a postoperative arterial blood sample obtained the first night after the surgical procedure. The occurrence and treatment of respiratory depression were recorded. For data analysis, respiratory depression was defined as either an arterial carbon dioxide tension of 55 mm Hg or more or a statement in the patient's chart that "respiratory depression" had occurred, without supporting arterial blood gas data. For statistical analysis, respiratory depression was categorized as follows: 0 = not present; 1 = present, as indicated by notation in the patient's chart without supporting blood gas data; 2 =present on the basis of blood gas data. Statistical Analysis.-Visual analog scales and verbal comfort level scores were analyzed statistically with use of the analysis of variance, rank sum test, and Fisher's exact
1033
Table I.-Summary of Demographic Data for Patients Who Underwent Radical Retropubic Prostatectomy, Stratified by Type of Analgesia* Factor Age (yrr] Weight (kg)'] Height (cm)t ASA physical statusj I 2 3
GroupE (N:: 33)
Group P (N = 33)
..68.7 ± 6.3 82.1±12.9 175.0 ± 7.2
65.9 ± 6.4 83.1 ± 9.1 176.1±7.1
No. % 4 12 24 73 5 15
No. % 3 9 25 76 5 15
*ASA = American Society of Anesthesiologists; group E:: epidural infusion of fentanyl for postoperative analgesia; group P = post-
operative intravenous' administration of morphine by patientcontrolled analgesia. tMean±SD. t1 :: healthy patient; 2 = mild systemic disease-no functional limitations; 3 = severe systemic disease---definite functional limitation.
test. The side effect profile was analyzed by using Fisher's exact test. The duration of hospital stay, use of narcotics, and all laboratory values were compared with use of the rank sum test. A two-tailed P value of 0.05 or less was considered statistically significant. RESULTS Of the 71 patients who entered the study, 5 were excluded from the data analysis for the following reasons: at the discretion of the attending anesthesiologist, 3 patients received a combination of epidural and general anesthesia intraoperatively; the PC A-morphine was changed to meperidine hydrochloride 12 hours postoperatively in 1 patient because of severe nausea and vomiting; and the epidural infusion was discontinued approximately 6 hours postoperatively in 1 patient because of a nursing shortage. As mentioned earlier, whether patients were in the general or epidural anesthesia stratum was found to have no statistically significant effect on the efficacy of postoperative analgesia relative to any of the determinations evaluated by analysis of variance. Therefore, for ease of presentation and for subsequent statistical analysis, patients were classified into two groups on the basis of their random assignment for postoperative analgesia to either epidural infusion of fentanyl (groups 1 and 2 = group E; N = 33) or PCAmorphine (groups 3 and 4 = group P; N = 33). Statistical analysis of these two groups showed no differences in age, weight, height or physical status, as defined by the American Society of Anesthesiologists (Table 1). Not all patients were available for assessment at all 11 observation times. In 19 patients, either the epidural or the PCA infusion was discontinued before 0800 hours on post-
Mayo Clin Proc, November 1992, Vol 67
1034 ANALGESIA AFTER RADICAL RETROPUBIC PROSTATECTOMY
operative day 3 for the following reasons: PCA-morphine was no longer being used or required (in 7 patients); resident's decision that epidural infusion was no longer required (3); increase in temperature (3 with epidural); mechanical difficulty with epidural catheter (occlusion in 1 and leak in 1); tenderness at epidural catheter site (1); development of transient paresthesia after a change of position in bed (1 with epidural); discontinuation of epidural infusion at request of surgical service (1); and dismissal from the hospital (1). In addition, at some observation times throughout the study period, the patients were not in their room or were sleeping. Furthermore, because of variations in the time of day when the patients returned from the operating room, fewer patients were assessed on the day of the surgical procedure than subsequently. Comfort Level During Resting.-The mean comfort level scores while the patients were resting in bed were lower (that is, greater comfort) in group E (epidural analgesia) than in group P (PCA) at every observation time (Fig. 1 and Table 2). At 9 of the 11 observation times, the difference was statistically significant (P::;;O.05). The percentage of patients who reported no pain during resting for each observation time was significantly higher in group E than in group P at all times except 2000 hours on postoperative day 2 and 0800 hours on postoperative day 3 (Table 2). Relief of Pain During Resting.-The mean scores for relief of pain during resting were lower (that is, better relief
100 80
"J!.
!If c:
60
~
40
Q.
20
Time
o
.0
1600 2000 0600 1200 1600 2000 loosJ L-P001--l
.0.1'3.3
0600 1200 1600 2000 0600 L-P002--l POD3
11IIIII3.4·6.7 1!Z16.8-10
Fig. 1. Distribution of comfort level scores during resting after radical retropubic prostatectomy. Vertical axis represents percentage of patients reporting. 0 =no pain; 10 =severe pain. DOS =day of surgery; E = epidural infusion of fentanyl; P = intravenous administrationof morphine by patient-controlled analgesia; POD = postoperative day. Asterisk = significant difference (P~0.05, rank sum test) in mean scores.
of pain) in group E than in group P at all times (Fig. 2 and Table 3). The differences were statistically significant (P::;;0.05) at 8 of the 11 observation times. Complete relief of pain during resting was achieved in 26% or more of the patients in group E at all observation
Table 2.-Distribution of Comfort Level Scores* During Resting After Radical Retropubic Prostatectomy, Stratified by Type of Analgesiat Day of operation 1600 2000
Postoperativeday 1 0800 1200 1600 2000
Postoperativeday 2 1200 1600 2000 0800
Postoperative day 3 0800
Epidural analgesia No. of patients Mean score SD No pain (%)
16 2.0 1.8 31.3
30 2.0 1.9 30.0
31 1.7 1.9 29.0
26 1.7 2.0 30.8
29 1.2 1.5 34.5
30 1.5 2.1 36.7
26 0.8 1.1 46.2
21 0.6 0.8 47.6
22 1.0 1.9 59.1
20 1.6 2.8 40.0
18 0.9 1.1 38.9
PCA-morphine No. of patients Mean score SD No pain (%)
14 5.0 2.4 0.0
29 3.3 1.7 0.0
29 3.3 1.9 3.4
28 2.7 1.6 3.6
29 2.5 1.8 6.9
27 2.0 1.5 11.1
29 2.1 1.7 3.4
25 1.6 1.0 8.0
21 2.0 1.7 14.3
20 1.9 1.9 20.0
21 1.3 1.2 14.3
0.002
0.008
0.003
0.019
0.002
0.034
PATRICK H. ALLAIRE, M.D., JOSEPH M. MESSICK, Jr., M.D., Department ofAnesthesiology; JOSEPH E. OESTERLING, M.D., Department of Urology; DAVID E. BYER, M.D., Department ofAnesthesiology; ROBERT P. MYERS, M.D., MICHAEL M. LIEBER, M.D., Department of Urology; ROBERT C. CHANTIGIAN, M.D., JEFFREY O. WELNA, D.O., Department ofAnesthesiology; DAVID E. PATTERSON, M.D., MICHAEL L. BLUTE, M.D., Department of Urology; KENNETH P. OFFORD, M.S., JENNIFER A. FERGUSON, B.S., Section ofBiostatistics
In a prospective, randomized study, continuous infusion of epidural fentanyl citrate (group E) was compared with patient-controlled intravenously administered morphine sulfate (group P) for analgesia in 66 men after radical retropubic prostatectomy. Although both methods provided satisfactory analgesia, the mean comfort level scores were lower (that is, greater comfort) in group E than in group P at all observation times. The difference in mean resting comfort level scores between groups E and P was statistically significant (P~O.05) at 9 of the 11 observation times. In addition, significant differences in comfort level scores were noted at 8 of the 11 observation times during deep breathing,S of 11 during coughing, and 3 of 9 during ambulation. Maximal and minimal comfort level scores recorded by each patient during the course of the study were significantly lower (that is, less pain) in group E than in group P for all four categories of activity. The percentage of patients who reported no pain was significantly higher in group E than in group P at 9 of 11 observation times during resting and 5 of 11 observation times during deep breathing. No significant differences were noted in side effect profiles or duration of hospital stay. In summary, when two effective methods of analgesia used after radical retropubic prostatectomy were compared prospectively, patients who received epidural infusion of fentanyl were more comfortable than those with patient-controlled intravenous administration of morphine, as evidenced by lower mean, maximal, and minimal comfort level scores and a greater proportion of patients with complete relief of pain.
Postoperative pain is unpleasant and associated with potentially detrimental physiologic effects. The intermittence of the analgesia associated with intramuscular or subcutaneous injections of narcotics'> has been attributed to fluctuating and unpredictable blood concentrations of these agents. The series of steps and attendant delays between the
Address reprint requests to Dr. J. M. Messick, Jr., Department of Anesthesiology, Mayo Clinic, Rochester, MN 55905. • Mayo Clin Proc 67:1031-1041,1992
patient's request for analgesia and the actual achievement of analgesia have been reviewed. 1 Recent advances in the management of postoperative pain have included on-demand patient-controlled analgesia (PCA)-usually accomplished by intravenous administration of a narcotic-and epidural infusion of a narcotic, a local anesthetic agent, or a combination of both.t' Perceived greater effectiveness of both intravenous peA and epidural infusion of narcotics for management of postoperative pain has led to their increased use for analgesia after various surgical procedures, including
1031
1032
ANALGESIA AFTER RADICAL RETROPUBIC PROSTATECTOMY
radical retropubic prostatectomy. Potential side effects of narcotics, including respiratory depression, may occur with any route of administration. A minimal risk of mechanical failure of the pump device with either PCA or epidural infusions and the need for a specially trained physician to accomplish epidural analgesia are other factors to be considered. A recent retrospective review of 76 patients who underwent radical retropubic prostatectomy indicated a significant decrease in duration of hospital stay postoperatively among patients who received a narcotic agent epidurally in comparison with those who received narcotics intramuscularly or intravenously (PCA).6 Prospective studies in which intravenously administered opioid PCA was compared with epidural infusion of an opioid in the treatment of pain after either an orthopedic surgical procedure or cesarean section have been reported.>'? We are unaware of any prospective randomized studies that have compared continuous epidural infusion of fentanyl citrate with intravenous administration of morphine sulfate by PCA in the treatment of pain after radical retropubic prostatectomy. The current investigation was designed to compare these two modalities of analgesia after this procedure at our institution in a randomized prospective study.
MATERIAL AND METHODS This study was approved by the Mayo Institutional Review Board. Seventy-one men scheduled to undergo bilateral pelvic lymph node dissection and radical retropubic prostatectomy for clinically localized prostatic cancer, who were judged by their anesthesiologists to be eligible to receive either morphine administered intravenously (PCA) or epidural infusion of fentanyl for postoperative analgesia, were randomized prospectively to receive one of these analgesic regimens. The surgical procedure performed on all patients was the same. Patients received either general anesthesia (thiopental sodium, isoflurane in a nitrous oxide-oxygen mixture, and intravenously administered morphine or fentanyl) or epidural anesthesia (lidocaine plus intravenous supplementation with fentanyl or midazolam hydrochloride [or both] as indicated). Epidural catheters were placed in the lumbar area (L2-3 or L3-4 interspace) by the staff anesthesiologist either preoperatively for intraoperative use or in the recovery room if the patient had received a general anesthetic agent and was to receive an epidural infusion of fentanyl postoperatively. Randomization proceeded in the following sequence. The anesthesiologist determined the patient's eligibility for participation, discussed the study, obtained informed consent, and decided in conjunction with the patient on either general or regional (epidural) anesthesia for the procedure. These two anesthesia groups then were considered as strata.
Mayo Clio Proc, November 1992, Vol 67
Within each stratum, patients were randomized to receive narcotics postoperatively by either epidural infusion or PCA such that after every six patients each group had three. After induction and stabilization of anesthesia, the anesthesiologist obtained the patient's random postoperative analgesia assignment. For discussion purposes, we defined four groups by a combination of strata and postoperative analgesia assignment. Group I patients (N =24) received general anesthesia intraoperatively and randomly assigned epidural infusion of fentanyl for postoperative analgesia. Group 2 patients (N = 11) received epidural anesthesia intraoperatively and randomly assigned epidural infusion of fentanyl for postoperative pain. Group 3 patients (N = 2S) received general anesthesia intraoperatively and randomly assigned intravenously administered morphine (hereafter referred to as PCA-morphine) for postoperative relief of pain. Group 4 patients (N = 11) received epidural anesthesia intraoperatively and randomly assigned PCA-morphine for postoperative analgesia; in these patients, the epidural catheters were removed in the recovery room. During statistical analyses, it became evident that choice of general or epidural anesthesia had no substantial effect on the efficacy of postoperative analgesia. Therefore, for further analysis, groups 1 and 2 were combined into group E (epidural), and groups 3 and 4 were combined into group P (PCA-morphine). For epidural analgesia, fentanyl was administered as a bolus (usually 100 ug), followed by a 1 ug/kg per h continuous infusion of fentanyl (10 ug/ml) in the recovery room. Postoperative management of these patients included nursing care based on guidelines developed for patients who receive epidural analgesia and monitoring by pulse oximetry for 12 hours after the patient returned to the hospital room. epidural infusions of fentanyl were managed by the Acute Pain Service in the Department of Anesthesiology. Patients in the PCA-morphine group received incremental 2- to 4-mg doses of morphine administered intravenously until comfortable in the recovery room, followed by intravenous infusion of morphine by a programmable pump in the hospital room. PCA orders were written by the primary surgical service and were not influenced or controlled by the Acute Pain Service. Typical orders were for a 2-mg bolus dose of morphine with a 10- to IS-minute lockout interval and a maximal dosage of 20 to 30 mg during a 4-hour period. Comfort levels of the patients, drug side effects and related observations, duration of hospital stay, and total dose of narcotics were compared between groups. Comfort Level of Patients.-Three visual analog scales were used to evaluate the patient's comfort. Each scale consisted of a lO-cm horizontal line on a 3- by S-inch (7.6by 12.7-cm) card. On one scale, patients were asked to evaluate their level of comfort at rest by placing a vertical
.11
Mayo Clin Proc, November 1992, Vol 67
ANALGESIA AFTER RADICAL RETROPUBIC PROSTATECTOMY
mark on an unmarked line, using a scale with no pain (0) at one end and the worst pain imaginable (10) at the other. On a second scale, patients were asked to represent the quality of reliefofpain by placing a vertical mark on an unmarked line designated "complete pain relief' (0) at one end and "no pain relief' (10) at the other. A third time-referenced scale asked patients to represent their comfort level relative to the first night after the operation by placing a vertical mark on a -5 to +5 scale, on which -5 ="much worse," 0 ="no change," and +5 = "much better." In addition, patients were asked to provide verbal ratings on a 0 to 10 scale (0 = no pain and 10 = worst pain imaginable) for their level of comfort when they took a deep breath, coughed, and ambulated. Patients were not asked to evaluate their pain while sitting because some surgeons preferred that their patients not sit in a straight chair for the first few days postoperatively. The presence or absence of pruritus, nausea, vomiting, and bladder spasm was scored on a 0 to 2 scale (0 = none present, 1 = present but not necessitating medication, and 2 = present and medication needed). Patients were assessed while awake at 1600 and 2000 hours on the day of the surgical procedure and at 0800, 1200, 1600, and 2000 hours daily until 0800 hours on the third postoperative day. In accordance with previous studies of postoperative analgesia and to allow our patients to sleep during the night, we did not evaluate pain scores after 2000 hours or before 0800 hours." Physical Features and RespiratoryData.-Age, weight, height, and physical status of the patients were recorded (Table I). In all patients, a preoperative arterial blood specimen was obtained for gas analysis during breathing of room air before induction of anesthesia in the operating suite. Oxygen saturation, arterial oxygen tension, arterial carbon dioxide tension, pH, base excess, sodium, hemoglobin, and creatinine measurements from the preoperative blood sample were compared with results from a postoperative arterial blood sample obtained the first night after the surgical procedure. The occurrence and treatment of respiratory depression were recorded. For data analysis, respiratory depression was defined as either an arterial carbon dioxide tension of 55 mm Hg or more or a statement in the patient's chart that "respiratory depression" had occurred, without supporting arterial blood gas data. For statistical analysis, respiratory depression was categorized as follows: 0 = not present; 1 = present, as indicated by notation in the patient's chart without supporting blood gas data; 2 =present on the basis of blood gas data. Statistical Analysis.-Visual analog scales and verbal comfort level scores were analyzed statistically with use of the analysis of variance, rank sum test, and Fisher's exact
1033
Table I.-Summary of Demographic Data for Patients Who Underwent Radical Retropubic Prostatectomy, Stratified by Type of Analgesia* Factor Age (yrr] Weight (kg)'] Height (cm)t ASA physical statusj I 2 3
GroupE (N:: 33)
Group P (N = 33)
..68.7 ± 6.3 82.1±12.9 175.0 ± 7.2
65.9 ± 6.4 83.1 ± 9.1 176.1±7.1
No. % 4 12 24 73 5 15
No. % 3 9 25 76 5 15
*ASA = American Society of Anesthesiologists; group E:: epidural infusion of fentanyl for postoperative analgesia; group P = post-
operative intravenous' administration of morphine by patientcontrolled analgesia. tMean±SD. t1 :: healthy patient; 2 = mild systemic disease-no functional limitations; 3 = severe systemic disease---definite functional limitation.
test. The side effect profile was analyzed by using Fisher's exact test. The duration of hospital stay, use of narcotics, and all laboratory values were compared with use of the rank sum test. A two-tailed P value of 0.05 or less was considered statistically significant. RESULTS Of the 71 patients who entered the study, 5 were excluded from the data analysis for the following reasons: at the discretion of the attending anesthesiologist, 3 patients received a combination of epidural and general anesthesia intraoperatively; the PC A-morphine was changed to meperidine hydrochloride 12 hours postoperatively in 1 patient because of severe nausea and vomiting; and the epidural infusion was discontinued approximately 6 hours postoperatively in 1 patient because of a nursing shortage. As mentioned earlier, whether patients were in the general or epidural anesthesia stratum was found to have no statistically significant effect on the efficacy of postoperative analgesia relative to any of the determinations evaluated by analysis of variance. Therefore, for ease of presentation and for subsequent statistical analysis, patients were classified into two groups on the basis of their random assignment for postoperative analgesia to either epidural infusion of fentanyl (groups 1 and 2 = group E; N = 33) or PCAmorphine (groups 3 and 4 = group P; N = 33). Statistical analysis of these two groups showed no differences in age, weight, height or physical status, as defined by the American Society of Anesthesiologists (Table 1). Not all patients were available for assessment at all 11 observation times. In 19 patients, either the epidural or the PCA infusion was discontinued before 0800 hours on post-
Mayo Clin Proc, November 1992, Vol 67
1034 ANALGESIA AFTER RADICAL RETROPUBIC PROSTATECTOMY
operative day 3 for the following reasons: PCA-morphine was no longer being used or required (in 7 patients); resident's decision that epidural infusion was no longer required (3); increase in temperature (3 with epidural); mechanical difficulty with epidural catheter (occlusion in 1 and leak in 1); tenderness at epidural catheter site (1); development of transient paresthesia after a change of position in bed (1 with epidural); discontinuation of epidural infusion at request of surgical service (1); and dismissal from the hospital (1). In addition, at some observation times throughout the study period, the patients were not in their room or were sleeping. Furthermore, because of variations in the time of day when the patients returned from the operating room, fewer patients were assessed on the day of the surgical procedure than subsequently. Comfort Level During Resting.-The mean comfort level scores while the patients were resting in bed were lower (that is, greater comfort) in group E (epidural analgesia) than in group P (PCA) at every observation time (Fig. 1 and Table 2). At 9 of the 11 observation times, the difference was statistically significant (P::;;O.05). The percentage of patients who reported no pain during resting for each observation time was significantly higher in group E than in group P at all times except 2000 hours on postoperative day 2 and 0800 hours on postoperative day 3 (Table 2). Relief of Pain During Resting.-The mean scores for relief of pain during resting were lower (that is, better relief
100 80
"J!.
!If c:
60
~
40
Q.
20
Time
o
.0
1600 2000 0600 1200 1600 2000 loosJ L-P001--l
.0.1'3.3
0600 1200 1600 2000 0600 L-P002--l POD3
11IIIII3.4·6.7 1!Z16.8-10
Fig. 1. Distribution of comfort level scores during resting after radical retropubic prostatectomy. Vertical axis represents percentage of patients reporting. 0 =no pain; 10 =severe pain. DOS =day of surgery; E = epidural infusion of fentanyl; P = intravenous administrationof morphine by patient-controlled analgesia; POD = postoperative day. Asterisk = significant difference (P~0.05, rank sum test) in mean scores.
of pain) in group E than in group P at all times (Fig. 2 and Table 3). The differences were statistically significant (P::;;0.05) at 8 of the 11 observation times. Complete relief of pain during resting was achieved in 26% or more of the patients in group E at all observation
Table 2.-Distribution of Comfort Level Scores* During Resting After Radical Retropubic Prostatectomy, Stratified by Type of Analgesiat Day of operation 1600 2000
Postoperativeday 1 0800 1200 1600 2000
Postoperativeday 2 1200 1600 2000 0800
Postoperative day 3 0800
Epidural analgesia No. of patients Mean score SD No pain (%)
16 2.0 1.8 31.3
30 2.0 1.9 30.0
31 1.7 1.9 29.0
26 1.7 2.0 30.8
29 1.2 1.5 34.5
30 1.5 2.1 36.7
26 0.8 1.1 46.2
21 0.6 0.8 47.6
22 1.0 1.9 59.1
20 1.6 2.8 40.0
18 0.9 1.1 38.9
PCA-morphine No. of patients Mean score SD No pain (%)
14 5.0 2.4 0.0
29 3.3 1.7 0.0
29 3.3 1.9 3.4
28 2.7 1.6 3.6
29 2.5 1.8 6.9
27 2.0 1.5 11.1
29 2.1 1.7 3.4
25 1.6 1.0 8.0
21 2.0 1.7 14.3
20 1.9 1.9 20.0
21 1.3 1.2 14.3
0.002
0.008
0.003
0.019
0.002
0.034
