Characterization of the analgesic actions of adenosine: comparison of adenosine and remifentanil infusions in patients undergoing major surgical procedures

Pain 101 (2003) 129–138 www.elsevier.com/locate/pain

Characterization of the analgesic actions of adenosine: comparison of adenosine and remifentanil infusions in patients undergoing major surgical procedures Atsuo F. Fukunaga a,*, George E. Alexander b, Charles W. Stark c a

Harbor-UCLA Medical Center, Department of Anesthesiology, 1000 West Carson Street, Torrance, CA, 90509, USA b Little Company of Mary Hospital, Department of Anesthesia, 4101 Torrance Blvd., Torrance, CA 90503, USA c Research and Education Institute, Harbor-UCLA Medical Center, 1124 West Carson Street, Torrance, CA, 90502, USA Received 13 February 2002; accepted 5 August 2002

Abstract Perioperative pain is still a major problem, and new pharmacological means should be explored to mitigate such pain. Adenosine is an ubiquitous endogenous substance; when exogenously administered, it provides a number of salutary effects including neuromodulation, antinociception, and cytoprotective actions. The aim of this study was to characterize the perioperative antinociceptive–analgesic effects of intraoperative adenosine infusion and determine the duration of actions in the postoperative period, and compare them to those of remifentanil in patients undergoing major surgical procedures in a double-blind study. Sixty-two patients were randomly assigned to one of the two treatments. After standard induction of anesthesia, the lungs were mechanically ventilated. Anesthesia was maintained with a constant alveolar concentration of inhaled anesthetics (3% desflurane and 65% nitrous oxide in oxygen). A variable-rate of intravenous infusion of adenosine (50–500 mg kg 21 min 21) or remifentanil (0.05–0.5 mg kg 21 min 21) was initiated 5 min before the skin incision and was titrated to maintain systolic blood pressure and heart rate within 20% of baseline values during surgery. Postoperative evaluations included the level of sedation, degree of pain severity, opioid analgesic (fentanyl, morphine) consumption, and cardiorespiratory variables for 48 h. Intraoperative inhibition of the cardiovascular responses to surgical stimulation could be equally achieved by adenosine or remifentanil, and both could maintain excellent hemodynamic stability. Postoperatively, however, there were striking differences: (1) initial pain score was reduced by 60% (P , 0:001) in the adenosine group compared to the remifentanil group and it remained lower throughout the 48 h recovery period; (2) postoperative morphine requirements during the first 0.25, 2 and 48 h were consistently lower in the adenosine group as compared to the remifentanil group (78, 71 and 42%, P , 0:001, respectively); (3) adenosine patients remained significantly less sedated at all evaluations; (4) the end-tidal and arterial carbon dioxide values in the remifentanil group were significantly higher when patients were admitted to the postanesthesia care unit. No adverse effect of adenosine was observed at any time. Intraoperative adenosine infusion provided a salutary recovery from anesthesia associated with a pronounced and sustained postoperative pain relief. Compared to remifentanil, adenosine significantly reduced the opioid requirements and minimized the side effects including protracted sedation, cardiorespiratory instability, nausea, and vomiting in the postoperative recovery period. q 2002 International Association for the Study of Pain. Published by Elsevier Science B.V. All rights reserved. Keywords: Postoperative pain; Adenosine; Remifentanil

1. Introduction Despite many advances in modern surgical and anesthetic care, perioperative pain relief has been and is still today one of the most compelling issues in the field of perioperative medicine, and has a significant impact on our health care

* Corresponding author. E-mail address: [email protected] (A.F. Fukunaga).

system (Utting and Smith, 1979). In an effort to improve perioperative pain treatment and hence improve recovery from anesthesia and surgery, new treatment modalities including preemptive analgesia (Woolf and Chong, 1993), multimodal pain therapy (Kehlet and Dahl, 1993), and new analgesic drugs (e.g. non-steroidal anti-inflammatory drugs a2-adrenergic agonists such as clonidine or dexmedetomidine, new opioids such as remifentanil) have been introduced. Constantly changing surgical stimulation of varying

0304-3959/02/$20.00 q 2002 International Association for the Study of Pain. Published by Elsevier Science B.V. All rights reserved. doi:10.1016/S0 304-3959(02)00 321-4

130

A.F. Fukunaga et al. / Pain 101 (2003) 129–138

degrees, surgical trauma and tissue damage initiate various types and intensities of nociceptive signals originating from the site of injury and these elicit an acute and complex array of neuroendocrine, hemodynamic, inflammatory, metabolic, and immunological changes. In addition, tissue injury causes a massive release of algogenic and proinflammatory mediators, which could further stimulate a series of inflammatory responses and thus sensitize both the peripheral and central nervous system. Traditionally, long-acting opioids have been used during anesthesia to blunt the pain, neuroendocrine and hemodynamic responses to surgical stimulation. Recently, a very potent and short-acting m-opioid receptor agonist that can effectively inhibit the surgical pain and stress-responses has been recognized as an intraoperative titratable analgesic drug (Glass, 1995). Adenosine is an ubiquitous endogenous purine nucleoside. Extracelluarly, adenosine has been shown to participate in numerous pathophysiological adaptation processes, including neuromodulatory (Williams, 1989), antiadrenergic (Perlini et al., 1998) and retaliatory actions (Newby, 1984), anti-inflammatory activities (Cronstein, 1997), and immune response processes (Priebe and Nelson, 1991). All these activities appear to be homeostatic and protective in nature. Of particular interest is adenosine’s antinociceptive properties (Sawynok, 1998 ). However, most of the research on the antinociceptive action in animal studies has been done using long-acting adenosine analogues because there has been a general perception that adenosine’s actions are short-lived due to its extremely short plasma half-life (Mo˝ ser et al., 1989). Therefore, the clinical use of adenosine itself has been mainly indicated for acute short-term treatments, such as for correcting paroxysmal supraventricular tachycardia (Adenocard) and for a diagnostic coronary artery vasodilation (Adenoscan). Clinically, we first observed a surprisingly long-lasting residual analgesia in the postoperative period after ATP infusion as a vasodilator to induce deliberate hypotension about two to three decades ago, but its antinociceptive– analgesic properties in surgical patients were not addressed. Numerous laboratory studies were subsequently carried out to demonstrate and characterize the antinociceptive action of intravenous (i.v.) adenosine and ATP (Fukunaga, 1997). Subsequent clinical investigations confirmed these findings in patients undergoing different types of surgical procedures (Fukunaga et al., 1995, Segerdahl et al., 1995, 1996, 1997, Zarate et al., 1999). However, adenosine-induced intraoperative antinociception and postoperative sustained analgesic activities were not clearly defined. The purpose of the present study was to characterize the perioperative antinociceptive activities activities of adenosine and determine the postinfusion recovery profiles in the postoperative period. In a randomized, double-blinded study, we compared the analgesic effects of i.v. adenosine with remifentanil in major orthopedic and gynecologic surgical procedures.

2. Material and methods 2.1. Study design and patients This study was approved by the Food and Drug Administration (FDA), the State of California Research Advisory Panel and the IRB, and all patients gave written informed consent. Sixty-two adult patients, 48.3 ^ 8.7 years, classified as American Society of Anesthesiologist (ASA) physical status I and II were scheduled for elective total abdominal hysterectomy (TAH, n ¼ 32), total knee arthroplasty (TKA, n ¼ 18) or total hip arthroplasty (THA, n ¼ 12). These major surgical procedures were selected because they were known to cause severe pain and stress elicited by massive surgical tissue injuries. Exclusion criteria included positive pregnancy test, history of asthma, gout, hepatic, renal and endocrine disorders, coronary artery disease, and ingestion of methylxanthine (caffeine, theophylline, cocoa) containing food or beverages within 12 h of surgery. Additional exclusionary criteria included patients who used illicit drugs, patients who had a known sensitivity to opiate or xanthine agonist or antagonist drugs (e.g. dipyridamole, aminophylline). Following FDA IND CMC procedures, the investigational drug service pharmacist randomized and prepared all study drugs. Blinded solutions of adenosine (5 mg/ml) and remifentanil (5 mg/ml) were prepared in sodium chloride (0.9%) and provided in 500 ml large volume parenteral (LVP) glass bottles 1. All other personnel participating in the surgery, anesthesia, subject evaluation, and data collection were blinded to the study drug assignment. 2.2. Intraoperative drug (adenosine or remifentanil) infusion First, all investigators participated in an open-label, randomized study of these two new drugs (remifentanil, n ¼ 10 or adenosine, n ¼ 11) to gain experience on intraoperative drug titration technique with the variable-rate of continuous infusion and to ascertain the estimated time of surgery and drug infusion volumes as well as the use of rescue pain medication during emergence from anesthesia. The remaining 41 patients were assigned in a double-blinded, randomized manner to either adenosine (n ¼ 20) or remifentanil (n ¼ 21). Standard anesthesia monitors including pulseoximeter, automated cuffed blood pressure (BP), electrocardiogram (EKG), peripheral forearm venous, and radial arterial catheters were placed. After premedication with i.v. midazolam (0.04 mg/kg), anesthesia was induced with i.v. fentanyl (1.5 mg/kg), i.v. lidocaine (50 mg), and i.v. propofol (1 mg/kg). Endotracheal intubation was facilitated with rocuronium (0.6 mg/kg) and general anesthesia was 1 Adenosine and remifentanil were reconstituted in a 1/1000 concentration ratio to ensure identical volume rate of infusion. Thus, the infusion rate adjustments could be made as the same number of millilitres per minute for both drugs that were clear in appearance.

A.F. Fukunaga et al. / Pain 101 (2003) 129–138

maintained with a constant end-tidal concentration of inhaled anesthetics (3% desflurane, and 65% nitrous oxide in oxygen) throughout the surgery. After attaining stable presurgical baseline values, the study drug infusion (either adenosine or remifentanil) was initiated 5 min before skin incision via an exclusive peripheral i.v. catheter. The study drug was administered by a variable-rate infusion method and titrated to inhibit the cardiovascular responses to surgical stimulation while maintaining systolic BP ^ 20% and heart rate (HR) ^ 20% from the baseline values throughout the surgical procedure and until the total content of the LVP was infused. The infusion rate ranges were 50– 500 mg kg 21 min 21 (total dose: 2500 mg) and 0.05– 0.5 mg kg 21 min 21 (total dose: 2.5 mg) for adenosine and remifentanil, respectively. The inhaled anesthetics were discontinued after the completion of skin closure. Upon emergence from anesthesia and after removal of the endotracheal tube, all patients were administered oxygen via a face mask. If pain was evidenced, i.v. 50 mg of fentanyl was administered immediately before the patient was transported to the postanesthesia care unit (PACU). If severe pain persisted, an additional 50 mg of fentanyl was repeated. Thereafter, i.v. morphine was administered on demand in PACU and followed by the patient controlled analgesia (PCA) in the ward. 2.3. Postoperative evaluation In the PACU, a blinded physician investigator continued the patient evaluation for at least 2 h and/or until the PACU discharge ($8: Aldrete post-anesthesia recovery score). Subsequent evaluations were recorded in the ward at 5,12, 24 and 48 h. Pain scores were evaluated after activity (coughing and deep breaths). Patients were instructed, a priori, on how to assess their pain on an 11-point verbal rating numeric (VRN) pain scale (DeLoach et al., 1998), defined as 0 ¼ no pain and 10 ¼ worst imaginable pain. They were also explained that they could safely self-administer morphine if required, and were instructed on how to used the PCA device (Lifecare 5500 PCA, Abbott Laboratories, Chicago IL). The PCA device was programmed to administer 1.5 mg i.v. boluses of morphine at minimal intervals of 6 min. While in PACU, patients were treated on demand with bolus of morphine (2–4 mg i.v.) at 2–5 min intervals for severe pain (.5 pain score) and repeated until the patient was comfortable, but not oversedated. Patient’s comfort, risk of overdosage and respiratory depression were continuously monitored. Level of sedation was assessed using a modified Ramsay sedation score (Ramsay et al., 1974), on a scale of 0–10: 0–1 ¼ awake and easily responsive; 2–3 ¼ cooperative, oriented, calm and easy to arouse; 4–5 ¼ response to commands and arousable, 6–7 ¼ asleep but responsive to loud auditory stimulus and somehow arousable, 8–9 ¼ asleep, sluggish response to loud auditory stimulus and difficult to arouse, 10 ¼ asleep, no response and not arousable. Orientation was defined as responding

131

correctly to all of the following cognitive questions: patient’s birthday, date of surgery, home address and a couple of simple arithmetic computation (e.g. 3 1 2 ¼ 5; 7 2 3 ¼ 4). Criteria for recovery from anesthesia and PACU discharge were assessed using a modified Aldrete recovery score (Aldrete, 1995). Physical behavior was assessed by the patients’ willingness or reluctance to cough, take deep breaths, move toes and sit in bed either with assistance or without assistance (i.e. exert some physical movement, which the investigator assessed as resistant, somewhat resistant or non-resistant). No prophylactic antiemetics or anti-shivering medications were administered in any group. Postoperative adverse effects including nausea, vomiting, shivering, pruritus and the patient’s discomfort behavior were assessed by direct observation in PACU and by interviewing and chart review afterwards. Patients and investigators subjectively assessed the satisfaction level with regard to the anesthesia and analgesia regimen as satisfied and that they would like to use the same treatment or unsatisfied and preferred an alternative treatment in future surgeries. 2.4. Statistical analysis A power analysis indicated that a minimum of 20 patients in each group would be required to detect a 30% difference in morphine consumption and verbal rating score for pain with a power of 0.8 (a ¼ 0:05). All data are expressed as the mean ^ SD, unless otherwise specified. Parametric data such as demographic and early recovery profiles were compared using the two-sample Student’s t-test. Non-parametric data such as the Aldrete score, opioid (morphine, fentanyl) consumption, pain score, and sedation score were analyzed using Mann–Whitney U test. Continuous variables are analyzed using analysis of variance, or the Kruskal–Wallis tests, depending on the distributions or significant differences in variances, followed by post-tests if indicated (Dunnett test for analysis of variance and Dunn test for Kruskal–Wallis test). Fisher’s Exact Test was used to compare the differences between groups for the incidence of nausea, vomiting, and shivering. P , 0:05 was considered significant. 3. Results 3.1. Subjects Sixty-two patients fulfilled the inclusion criteria. Patients were well balanced with regard to demographic characteristics (Table 1). Pre-existing medical conditions, type of surgical procedures, duration of surgery, anesthesia, and drug infusion were all comparable in both groups. Of these, the preliminary 21 patients were enrolled in an open-label, randomized study in order to optimize the research protocol. Although the results of the open-label study were strikingly similar to those of the blinded ones,

132

A.F. Fukunaga et al. / Pain 101 (2003) 129–138

Table 1 Demographic, vitals and time characteristics for the two treatment groups a Study group

Remifentanil (n ¼ 21)

Adenosine (n ¼ 20)

Sex (male/female) ASA physical status I/II Age (year) Weight (kg) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Heart rate (bpm) Body temperature ( oC) Anesthesia time (min) Surgery time (min) Drug Infusion time (min)

4/17 3/18 46 ^ 9 85 ^ 19 137 ^ 12 77 ^ 13 81 ^ 13 36.7 ^ 0.3 208 ^ 70 157 ^ 61 143 ^ 49

4/16 5/15 51 ^ 8 70 ^ 10 134 ^ 14 80 ^ 9 80 ^ 12 36.6 ^ 0.4 230 ^ 50 182 ^ 48 162 ^ 36

a

resume spontaneous ventilation in both groups. No patient in the adenosine group showed any evidence of cardiorespiratory depression during or after recovery from anesthesia (Table 2 ). However, in the remifentanil group, there was a tendency of higher BP and HR that required the use of a and b blockers in some patients. Also, a moderate degree of respiratory depression persisted postsurgery as evidenced by both the end-tidal gas and the significantly elevated arterial blood PaCO2 values compared to adenosine (51 ^ 7 vs. 42 ^ 6 mmHg, respectively, P , 0:05). The higher PaCO2 was particularly prominent in the first 2 h after surgery. 3.3. Emergence and recovery from anesthesia

Values are mean ^ SD, * P , 0:05.

they were not included for the statistical analysis. The remaining 41 patients were studied, in a double-blinded, randomized protocol, to either treatment with adenosine [TAH (n ¼ 11), TKA (n ¼ 6), THA (n ¼ 3)] or remifentanil [TAH (n ¼ 12), TKA (n ¼ 6), THA (n ¼ 3)]. Three patients were regular users of illicit drugs, but had concealed that information at the time of enrollment and throughout the postoperative evaluation. The patients disclosed that information to the investigator physician just before they were discharged from the hospital. When the concealed study was uncovered at the completion of the study, it was discovered that all addicted patients happened to be in the adenosine group. Since the individual analysis of each patient was remarkably similar to the rest of the patients, these data were not excluded. 3.2. Postoperative cardiopulmonary profiles At the end of surgery, and after reversal of the residual neuromuscular blocking drug effect, all patients were able to

After discontinuation of the inhaled anesthetics, all patients awoke quickly. They responded to verbal command and breathed spontaneously, thus the tracheal tube could be safely removed within 5 min in both groups. However, the times for complete orientation was significantly faster in the adenosine group compared to remifentanil group (6 ^ 2 vs. 31 ^ 21 min, P , 0:05). 3.4. Adverse events Overall incidences of postoperative nausea and vomiting were 48 and 29% in the remifentanil group and 20 and 15% in the adenosine group, respectively (Table 3). It was not until 12 h postsurgery that the incidences of nausea and vomiting declined in the remifentanil group whereas it was nil after 2 h in the adenosine group. No patient in either group complained of pruritus. 3.5. Postoperative pain and opioid requirement Awakening from anesthesia was associated with severe pain in the remifentanil treated patients, which all (100%) required immediate analgesic treatment with an initial i.v. bolus fentanyl (50 mg) and a subsequent 50 mg additional

Table 2 Postoperative hemodynamic data a Postoperative time (h) Study drug

Vital signs

R A

HR (bpm)

R A

Preoperative control

0.25

1

2

5

12

24

48

79 ^ 11 85 ^ 12

81 ^ 16 92 ^ 17

77 ^ 16 89 ^ 15

78 ^ 15 93 ^ 11

87 ^ 15 89 ^ 12

88 ^ 18 93 ^ 14

93 ^ 18 97 ^ 15

92 ^ 14 93 ^ 13

SBP (mmHg)

140 ^ 17 140 ^ 18

138 ^ 17 125 ^ 25

135 ^ 20 122 ^ 22

132 ^ 21 121 ^ 22

129 ^ 20 125 ^ 22

123 ^ 21 121 ^ 21

133 ^ 17 127 ^ 19

134 ^ 12 125 ^ 16

R A

DBP (mmHg)

79 ^ 15 77 ^ 11

77 ^ 12 67 ^ 14

70 ^ 9 68 ^ 13

70 ^ 10 65 ^ 13

75 ^ 12 71 ^ 13

69 ^ 11 69 ^ 12

74 ^ 12 71 ^ 12

71 ^ 10 71 ^ 8

R A

MAP (mmHg)

99 ^ 15 98 ^ 12

97 ^ 13 86 ^ 17

92 ^ 12 86 ^ 13

91 ^ 12 84 ^ 14

93 ^ 13 89 ^ 14

87 ^ 14 86 ^ 14

94 ^ 12 89 ^ 12

92 ^ 11 89 ^ 10

R A

BT ( oC)

36.3 ^ 0.4 36.4 ^ 0.7

35.6 ^ 0.5 36.1 ^ 1.0

36.1 ^ 0.6 36.4 ^ 0.3

35.9 ^ 0.4 36.7 ^ 0.3

36.6 ^ 0.5 36.8 ^ 0.8

37.2 ^ 0.6 37.4 ^ 0.4

37.5 ^ 0.6 37.7 ^ 0.6

38.0 ^ 0.7 37.8 ^ 0.5

a

Values are mean ^ SD; R, Remifentanil (n ¼ 21); A, Adenosine (n ¼ 20); HR, heart rate; SBP, systolic blood pressure; DBP, diastolic blood pressure; MAP, mean arterial pressure; BT, body temperature.

A.F. Fukunaga et al. / Pain 101 (2003) 129–138

133

Table 3 Postoperative recovery and side effects profile in the two treatment groups a Postoperative times (h) Study drugs (% overall) 0.25

1

2

5

12

24

48

% Nausea

R (48%) A (20%)

29 10

19 5

29 5

29 0*

33 0*

19 0

0 0

% Vomiting

R (29%) A (15%)

5 5

0 0

5 10

10 0

14 0

10 0

0 0

% Shivering

R (14%) A (0%)

14 0

5 0

0 0

0 0

0 0

0 0

0 0

Physical behavior score (0–3)

R A

0.0 ^ 0.0 0.1 ^ 0.2

0.0 ^ 0.0 0.1 ^ 0.2

0.0 ^ 0.0 0.1 ^ 0.2

0.0 ^ 0.2 0.5 ^ 0.5*

0.2 ^ 0.4 0.9 ^ 0.6*

1.3 ^ 0.9 1.8 ^ 0.9

2.1 ^ 0.7 2.1 ^ 0.6

Aldrete Score (0–10)

R A

6.7 ^ 1.3 8.3 ^ 1.0*

7.5 ^ 1.4 8.9 ^ 1.0*

7.6 ^ 1.4 9.3 ^ 0.8*

8.9 ^ 1.2 9.8 ^ 0.6*

9.5 ^ 0.8 10.0 ^ 0.0*

10.0 ^ 0.0 10.0 ^ 0.0

10.0 ^ 0.0 10.0 ^ 0.0

a

R, Remifentanil; A, Adenosine. Values are incidence (%), mean ^ SD. * P , 0:05.

administration (total 100 mg). In contrast, 60% of patients in the adenosine group received 50 mg followed by 50 mg, and 25% required no fentanyl at all. The total fentanyl required by the remifentanil patients was 100 ^ 0 vs. 68 ^ 42 mg in adenosine patients, respectively, P , 0:05. Upon arrival in PACU, the initial pain score was significantly lower in the adenosine group compared to the remifentanil group (3.6 ^ 3.3 vs. 9.0 ^ 1.7, respectively, P , 0:001), and this lower pain scores remained significantly low for the next 48 h (Fig. 1). Furthermore, cumulative morphine consumption during the first 2 h in PACU was significantly lower

(71%) compared to remifentanil group (7 ^ 6 vs. 24 ^ 8 mg, P , 0:001, respectively) and remained lower thereafter. The total cumulative morphine consumption in the first 48 h was significantly lower (42%) in the adenosine than in the remifentanil group (53 ^ 26 vs. 92 ^ 35 mg, P , 0:001, respectively) (Fig. 2). 3.6. Postoperative sedation and physical behavior As indicated by the modified Ramsay sedation score data (Fig. 3), the adenosine-treated patients were significantly

Fig. 1. Postoperative pain score in adenosine (solid histogram) or remifentanil (plain histogram) group. Each patient was asked to assess his or her pain on an eleven-point VRN pain scale as they had been instructed preoperatively with zero for no pain to ten for worst imaginable pain. Results are expressed as mean ^ SEM. Differences between groups at the time indicated. ** P , :01, *** P , 0:001.

134

A.F. Fukunaga et al. / Pain 101 (2003) 129–138

Fig. 2. Postoperative cumulative morphine consumption in adenosine (solid histogram) or remifentanil (plain histogram) group. Patients were given i.v. bolus morphine on demand in PACU (,2 h), followed by PCA infusion pump administration. Each histogram represents mean ^ SEM at each designated time. Since the morphine effect is longitudinally cumulative, differences between groups are compared at a specific time (upon arrival in the PACU , 0.25 h, discharge from the PACU 2 h, and the following 2 days 48 h) to minimize the statistical bias. ** P , 0:01, ***P , 0:001.

less sedated and remained more alert compared to the remifentanil group during the entire postoperative course. All adenosine patients were non-resistant to exert some physical movement but the remifentanil patients were resistant or somehow resistant. No patient required the use of an antagonist, either naloxone or aminophylline, to treat or reverse the occurrence of any undesired complication. No patient was discontinued from the study for adverse effects. After completion of the

48 h study period, both the patients and researchers rated the subjective overall satisfaction to be superior with adenosine compared to remifentanil.

4. Discussion The results of the present study confirm that intraoperative adenosine infusion resulted in pronounced and sustained

Fig. 3. Postoperative sedation level in adenosine group (solid histogram) or remifentanil (plain histogram) group. Level of sedation was assessed by using modified Ramsay sedation scores on a scale of 0–10: 0–1 ¼ awake and easily responsive; 2–3 ¼ cooperative, oriented, and calm, easy to arouse; 4–5 ¼ response to commands, arousal, 6–7 ¼ asleep but response to loud auditory stimulus, somehow aroused, 8–9 ¼ asleep, sluggish response to loud auditory stimulus, difficult to arouse 10 ¼ asleep, not aroused. Results are expressed as mean ^ SEM. Differences between groups at the time indicated. NS ¼ nonsignificant. * P , 0:05, ** P , 0:01, *** P , 0:001.

A.F. Fukunaga et al. / Pain 101 (2003) 129–138

postoperative pain relief without causing excessive sedation or cardiorespiratory depression. Although both adenosine and remifentanil are known to be quickly eliminated from the circulating plasma, their intraoperative administration resulted in contrasting postoperative recovery profiles. The marked adenosine-induced postoperative analgesic activity, which persisted for over 48 h, clearly exceeded the expected effect at its plasma level, since circulating adenosine’s plasma half-life is less than 10 s (Mo¨ ser et al., 1989). Indeed, the rapid turnover of adenosine in the plasma under the influences of cellular uptake and/or metabolism by vascular endothelium, erythrocyte, and adenosine deaminase makes it difficult to estimate the circulating plasma levels of adenosine in vivo (Dawicki et al., 1985). Therefore, a traditional pharmacological notion based on the pharmacokinetic and pharmacodynamic theory does not appear to fully explain the adenosine-induced prolonged analgesic activity that lasted for a considerable period of time after the infusion had stopped. It is possible that a number of mechanisms of action may be responsible for its long lasting effect. After intrathecal adenosine injection, prolonged alleviation of hypersensitivity and increase in intracellular concentrations of adenosine in intrinsic spinal cord neurons have been shown in a neuropathic pain model despite its short residence time in the cerebrospinal fluid (Bantel et al., 2002) It could be argued that comparing the analgesic effects with those of remifentanil may be limited in light of the conflicting reports suggesting that remifentanil can produce acute tolerance (Vinik and Kissin, 1998; Guignard et al., 2000). However, no clinical evidence of acute opioid tolerance after intraoperative infusion of remifentanil has been found by other investigators (Cortinez et al., 2001) and the notion of acute development of tolerance to opioids in humans remains controversial (Schraag et al., 1999; Gustorff et al., 2002). In the present study, we chose remifentanil as the comparative-control group because of its similarities to adenosine with respect to: (a) their pharmacokinetics (extremely short plasma elimination time); (b) their speed and effectiveness in suppressing the acute cardiovascular responses to surgical stimulation; (c) both drugs can be similarly administered by titration utilizing the variable-rate infusion technique, and thereby, most suitable for a double-blinded, comparative study protocol. Furthermore, double-blinded, placebo (saline)-control studies for analgesic efficacy of intraoperative adenosine infusions were previously reported in clinical studies of shoulder surgery (30 patients), breast surgery (72 patients), and abdominal hysterectomy (41 patients) (Segerdahl et al., 1995, 1996, 1997, respectively). The presumably intraoperative antinociceptive effect of adenosine compared to that of remifentanil in our study showed that both drugs provided similarly effective suppression of cardiovascular responses to surgical stimulation and maintained excellent hemodynamic stability during surgery. Postoperatively, however, patients in the remifentanil group experienced high incidence of stormy emergence associated with pain,

135

which required immediate and proper treatment with i.v. rescue medications. These results are in agreement with previous reports (Dershwitz et al., 1995). The cumulative morphine consumption in the first 24 h in our remifentanil group was comparable to the morphine consumptions of other studies for TAH (Blackburn et al., 1995; Gan et al., 1997) and total hip or knee arthroplasty (Etches et al., 1995). This supports the notion that the pain experienced by our patients was reasonably standard. Despite initial higher opioid consumption and continued additional selfadministered PCA morphine, patients in the remifentanil group did not experience much improved pain relief. In contrast, intraoperative adenosine appeared to have conferred preemptive analgesia. It was intriguing that 5– 10 min after the inhaled anesthetics were discontinued, all adenosine-treated patients were awake and clear-headed, but experienced minimal pain, and more than 40% of these patients requested little or no morphine in the first 2 h of PACU observation. They remained calm and tranquil, but not oversedated as assessed by their positive responses to verbal commands and cognitive questions. They were also less resistant to cough, and were able to take deep breaths, move their toes or sit in bed without assistance. These behavioral responses correlated well with the pain scores and morphine consumptions. Since patients receiving adenosine were less sedated and were more alert, it is conceivable that these patients were more susceptible to pain and discomfort after surgery, and thus they would self-administer more morphine. However, the cumulative morphine consumption in the first 2 h and up to 48 h postsurgery in the adenosine group was significantly lower (71 and 42% reduction, respectively) compared to the remifentanil group. The lower opioid consumption may have accounted in part for the lower incidence of postoperative nausea and vomiting. Several studies (Segerdahl et al., 1995, 1996, 1997; Za´ rate et al., 1999) previously reported an analgesic effect for perioperative adenosine infusion with similar results, but appeared less conclusive given the methods and lower doses of adenosine infusion. For example, in patients undergoing abdominal hysterectomy, Segerdahl et al. (1997), in a double-blind placebo controlled study, found no difference in postoperative visual analogue scale (VAS) pain scores between the groups. On the other hand, Za´ rate et al. (1999), in a comparative study with remifentanil, demonstrated significantly lower VAS pain scores with adenosine, which persisted for 12 h but were not significant at 24 h. Both studies reported a significant reduction of opioid analgesic requirement (18 and 27%) in the first 24 h postoperative period, respectively. In contrast, our results clearly demonstrated that initial VRN pain scores as well as the opioid consumptions were both significantly lower in the adenosine group, which remained consistently low throughout the 48 h study period. Our adenosine hysterectomy patients showed a significant reduction in both VRN pain score (60%) and opioid consumption (46%) at 24 h postoperative assessment.

136

A.F. Fukunaga et al. / Pain 101 (2003) 129–138

The differences in the severity of postoperative pain experience and opioid consumption between our results and those of the above studies are probably methodological. We evaluated the degree of pain intensity during activity (deep breathing and coughing) and chose a 11-point VRN pain scale over the VAS pain scale in order to minimize the difficulties associated with the VAS pain score, as patients appear to have difficulty in recognizing the VAS scale that is associated with blurred vision due to postoperative residual anesthesia and/or use of high-dose opioids. In addition, the mode of intraoperative infusion and the total dose of adenosine administered may have further influenced the above results. Segerdahl et al. (1997) administered adenosine at a constant low infusion rate of 80 mg kg 21 min 21 (estimated mean total dose: 950 mg) while varying the inspired concentration of isoflurane during surgery. Za´ rate et al. (1999) administered higher dose of adenosine infusion (166 ^ 17 mg kg 21 min 21, mean total dose: 1400 ^ 164 mg) in combination with desflurane and nitrous oxide, both of which were adjusted at a variable-rate according to the targeted ranges of the bispectral index values. In contrast, we maintained a constant end-tidal (alveolar) concentration of inhaled anesthetics during surgery and administered higher-dose of adenosine infusion, which was titrated in response to surgical stimulation to maintain hemodynamic stability throughout the surgical procedure (292 ^ 82 mg kg 21 min 21, total dose: 2500 mg). It is noteworthy that the prolonged analgesic activity in the above studies and the present study is consistent, and appears to indicate a dose–response effect. However, a more systematic study may be required to confirm such a correlation. It is well recognized that adenosine plays a multifaceted and a complex, but significant role in the perception of pain at both central and peripheral sites. The complexity is further multiplied because of the different actions at multiple sites within the pain signaling and transmission as well as perception systems. Four subtypes of adenosine receptors have been identified (Olah and Stiles, 1995). These receptors are classified on the basis of their inhibitory or stimulatory effects via G protein on adenylyl cyclase. The A1- and A3-subtypes are coupled to the inhibitory pertussis toxin-sensitive G protein, while the A2A- and A2B-receptors are coupled in a high- and low-affinity manner, respectively, to the stimulatory G protein. A number of these can influence pain signaling, with the effect depending on receptor subtype, affinity to adenosine concentration and localization of the receptors. At peripheral nerve terminals, A1 receptors produce inhibition of pain transmission (antinociceptive), while A2A and A3 receptors facilitate pain transmission (pronociceptive). In the spinal cord, there is convincing evidence for a key role of the adenosine A1 receptor in the central control of nociception at the spinal cord level inhibiting pain (antinociceptive), but the role of A2A, A2B and A3 receptors at this site are not clearly defined (Sawynok, 1998). Pharmacological evidence has been provided for a close functional relationship between adenosine A1 and NMDA

receptors in the cerebral cortex (Hoehn and White, 1990) and in the spinal cord (Reeve and Dickenson, 1995). Autoradiographic studies have revealed the highest density of adenosine receptors in the substantia gelatinosa of the spinal cord dorsal horn (Choca et al., 1988), which serve as important relays of nociceptive transmission. A number of electrophysiological studies have shown that adenosine acts both presynaptically to reduce neurotransmitter release, and postsynaptically to hyperpolarize the spinal cord neurons of the substantia gelatinosa by interactions with ATP-sensitive-K 1 channels to increase conductance, so as to inhibit primary afferent synaptic transmission from dorsal root fibers modulating somatosensory functions (Salter et al., 1993; Li and Perl, 1994). This action most likely accounts for inhibition of the postsynaptic actions of substance P and excitatory amino acids. The latter action may be of particular relevance to the inhibition of ‘wind up’ phenomena that has been observed in the spinal cord following application of an adenosine A1 analogue (Reeve and Dickenson, 1995). Furthermore, it has been demonstrated that activation of spinal A1 receptor suppresses the noxiously evoked activity in dorsal horn wide dynamic range neurons in response to somatic (Sumida et al., 1998) as well as visceral (Smith et al., 1996) stimulation in both sensitized and non-sensitized spinal cords as demonstrated by spinally administered A1 receptor agonist, lphenyl-isopropyl adenosine. More specifically, it has been shown that A1 receptor agonists inhibit spinal sensory transmission related to nociception (Nakamura et al., 1997). Recent findings have shown a role of spinal A1 receptor involvement of antinociceptive action in inflammatory pain processes (Honore et al., 1998). It has been suggested that spinal A1 receptors may control acute and more persistent inflammatory nociceptive responses of dorsal horn neurons (Reeve and Dickenson, 1995), as evidenced by the inhibition of hyperalgesia on the formalin test (Malmberg and Yaksh, 1993), and the carrageenan-induced rat inflammatory and thermal hyperalgesia model (Poon and Sawynok, 1998). Such observations have been recently corroborated in human studies with spinal administration of adenosine (Eisenach et al., 2002). Collectively, these observations suggest that adenosine produces long-lasting analgesic actions by central, particularly spinal actions mediated via A1 receptors. In addition, it is now well recognized that the severity of postoperative pain may also be attributed not only to central mechanisms but also to peripheral inflammatory processes including peripheral nociceptor sensitization. In the broader perspective, postoperative pain is probably the prototypic inflammatory reaction to surgical tissue damage wherein the release of numerous proinflammatory mediators, including various proinflammatory cytokines has been evidenced in surgical patients (Naito et al., 1992; Kato et al., 1997). Furthermore, activation of adenosine receptors on the peripheral tissues and immune cells are variably expressed depending on species, cell types and local adenosine

A.F. Fukunaga et al. / Pain 101 (2003) 129–138

concentrations. Although in some instances, activation of A1, A2B, or A3 receptors in the peripheral tissues can produce proinflammatory responses, selective stimulation of A2A or A3 receptors on immune cells produces a series of responses that can be categorized as anti-inflammatory. These include decreased neutrophil adherence to biological substrates (Cronstein, 1994, 1997; Sullivan and Linden, 1998) and diminished release of tissue destructive oxidative (Cronstein, 1994) and non-oxidative products (Bouma et al., 1994). Peripherally circulating adenosine in humans may also act via A2A and A3 receptors to inhibit neutrophil degranulation and add to the anti-inflammatory potential in neutrophil-mediated tissue injury (Bouma et al., 1997). Recently, it was demonstrated that activation of adenosine A2 and A3 receptors inhibits the production of tumor necrosis factor-a (TNF-a) of human polymorphonuclear leukocytes (Thiel and Chouker, 1995), and macrophages (Sajjadi et al., 1996) responsible for some of the deleterious inflammatory responses. Moreover, adenosine also inhibits the production of several proinflammatory cytokines (IL-1, IL-6, IL-8 and TNF-a) by lipopolysaccharied-activated human monocytes (Bouma et al., 1994). Thus, it is conceivable that the involvement of the A2A and A3 receptors mediated peripheral anti-inflammatory effects could have been another important contributing factor for the pronounced postoperative pain relief observed in our study. Recent understanding of the mechanisms of perioperative pain indicate that during surgical tissue injury and subsequent inflammatory processes, both central and peripheral sensitization must be prevented to achieve effective postoperative pain relief (Woolf and Chong, 1993). Our results suggest that intraoperative adenosine may have provided concerted action of inhibitory surgical nociceptive transmission via central A1 receptor-mediated antinociception and prevention of peripheral inflammatory processes via A2A and possibly A3 receptors-mediated anti-inflammatory effects that preempted the establishment of pain hypersensitivity during and after surgery. Further studies, however, are warranted to better elucidate the main underlying mechanisms of adenosine’s analgesic action in surgical patients. In conclusion, this randomized, double-blinded clinical study demonstrated that the use of intraoperative adenosine infusion provided a salutary perioperative analgesia and recovery from anesthesia associated with a pronounced and sustained postoperative pain relief. Compared to remifentanil, adenosine significantly reduced perioperative opioid requirements, and minimized side effects, including protracted sedation, cardiorespiratory depression, nausea and vomiting in the postoperative recovery period.

Acknowledgements The authors thank Kyowa Hakko Inc. for providing adenosine, and the patients who participated in this study. We also wish to thank Drs K. Fukuda and H. Sadeghi for

137

helping with the illustrations and Dr J.S. McDonald for editorial assistance.

References Aldrete JA. The post-anesthesia recovery score revisited. J Clin Anesth 1995;7:89–91. Bantel C, Tobin JR, Li X, Childers SR, Chen SR, Eisenach JC. Intrathecal adenosine following spinal nerve ligation in rat: short residence time in cerebrospinal fluid and no change in A1 receptor binding. Anesthesiology 2002;96:103–108. Blackburn A, Stevens JD, Wheatley RG, Madej TH, Hunter D. Balanced analgesia with intravenous ketorolac and patient-controlled morphine following lower abdominal surgery. J Clin Anesth 1995;7:103–108. Bouma MG, Stad RK, van den Wildenberg FAJM, Buurman WA. Differential regulatory effects of adenosine on cytokine release by activated human monocytes. J Immunol 1994;153:4159–4168. Bouma MG, Jeunhomme TMMA, Boyle DL, Dentener MA, Voitenok NN, van den Wildenberg FAJM, Buurman WA. Adenosine inhibits neutrophil degranulation in activated human whole blood. Involvement of adenosine A2 and A3 receptors. J Immunol 1997;158:5400–5408. Choca JI, Green RD, Proudfit HK. Adenosine A1 and A2 receptors of the substantia gelatinosa are located predominantly on intrinsic neurons: an autoradiography study. J Pharmacol Exp Ther 1988;247:757–764. Cortinez LI, Brandes V, Munoz HR, Guerrero ME, Mur M. No clinical evidence of acute opioid tolerance after remifentanil-based anaesthesia. Br J Anaesth 2001;87:866–869. Cronstein BN. Adenosine, an endogenous anti-inflammatory agent. J Appl Physiol 1994;76:5–13. Cronstein BN. Adenosine regulation of neutrophil function and inhibition of inflammation via adenosine receptors. In: Jacobson KA, Jarvis MF, editors. Purinergic approaches in experimental therapeutics, New York, NY: Wiley-Liss, 1997. pp. 285–299. Dawicki DD, Agarwal KC, Parks Jr RE. Role of adenosine uptake and metabolism by blood cells in the antiplatelet actions of dipyridamole, dilazep and nitrobenzylthioinosine. Biochem Pharmacol 1985;34:3965–3972. DeLoach LJ, Higgins MS, Caplan AB, Stiff JL. The visual analog scale in the immediate postoperative period: intrasubject variability and correlation with a numeric scale. Anesth Analg 1998;86:102–106. Dershwitz M, Randel GI, Rosow CE, Fragen RJ, Connors PM, Librojo ES, Shaw DL, Peng AW, Jamerson BD. Initial clinical experience with remifentanil, a new opioid metabolized by esterases. Anesth Analg 1995;81:619–623. Eisenach JC, Hood DD, Curry R. Preliminary efficacy assessment of intrathecal injection of an American formulation of adenosine in humans. Anesthesiology 2002;96:29–34. Etches RC, Warriner CB, Badner N, Buckley DN, Beattie WS, Chan VWS, Parsons D, Girad M. Continuous intravenous administration of ketorolac reduces pain and morphine consumption after total hip or knee arthroplasty. Anesth Analg 1995;81:1175–1180. Fukunaga AF, Miyamoto TA, Kikuta Y, Kaneko Y, Ichinohe T. Role of adenosine and adenosine triphosphate as anesthetic adjuvants. In: Belardinelli L, Pelleg A, editors. Adenosine and adenine nucleotides: from molecular biology to integrative physiology, Boston, MA: Kluwer Academic, 1995. pp. 511–523. Fukunaga AF. Adenosine compounds. In: White PF, editor. Textbook of intravenous anesthesia, Baltimore, MD: Williams and Wilkins, 1997. pp. 413–432. Gan TJ, Ginsberg B, Glass PSA, Fortney J, Jhaveri R, Perno R. Opioidsparing effects of a low-dose infusion of naloxone in patient-administered morphine sulfate. Anesthesiology 1997;87:1075–1081. Glass PSA. Remifentanil: a new opioid. J Clin Anesth 1995;7:558–563. Guignard B, Bossard AE, Coste C, Sessler DI, Lebrault C, Alfonsi P, Fletcher D, Chauvin M. Acute opioid tolerance. Intraoperative remifen-

138

A.F. Fukunaga et al. / Pain 101 (2003) 129–138

tanil increases postoperative pain and morphine requirement. Anesthesiology 2000;93:409–417. Gustorff B, Nahlik G, Hoerauf KH, Kress HG. The absence of acute tolerance during remifentanil infusion in volunteers. Anesth Analg 2002;94:1223–1228. Hoehn K, White TD. Role of excitatory amino acid receptors in K 1- and glutamate-evoked release of endogenous adenosine from rat cortical slices. J Neurochem 1990;54:256–265. Honore P, Buritova J, Chapman V, Besson JM. UP 202-56, an adenosine analogue, selectively acts via A1 receptors to significantly decrease noxiously-evoked spinal c-Fos protein expression. Pain 1998;75:281– 293. Kato M, Suzuki H, Murakami M, Akama M, Matsukawa S, Hashimoto Y. Elevated plasma levels of interleukin-6, interleukin-8, and granulocyte colony-stimulating factor during and after major abdominal surgery. J Clin Anesth 1997;9:293–298. Kehlet H, Dahl JB. The value of ‘multimodal’ or ‘balanced analgesia’ in postoperative pain treatment. Anesth Analg 1993;77:1048–1056. Li J, Perl ER. Adenosine inhibition of synaptic transmission in the substantia gelatinosa. J Neurophysiol 1994;72:1611–1621. Malmberg AB, Yaksh TL. Pharmacology of the spinal action of ketorolac, morphine, ST-91, U50488H, and L-PIA on the formalin test and an isobolographic analysis of the NSAID interaction. Anesthesiology 1993;79:270–281. Mo¨ ser GH, Schrader J, Deussen A. Turnover of adenosine in plasma of human and dog blood. Am J Physiol 1989;256(Cell Physiol 25):C799– 806. Naito Y, Tamai S, Shingu K, Shindo K, Matusi T, Segawa H, Nakai Y, Mori K. Responses of plasma adrenocorticotropic hormone, cortisol, and cytokines during and after upper abdominal surgery. Anesthesiology 1992;77:426–431. Nakamura I, Ohta Y, Kemmotsu O. Characterization of adenosine receptors mediating spinal sensory transmission related to nociceptive information in the rat. Anesthesiology 1997;87:577–584. Newby AC. Adenosine and the concept of ‘retaliatory metabolites’. Trends Biochem Sci 1984;9:42–44. Olah ME, Stiles GL. Adenosine receptor subtypes: characerization and therapeutic regulation. Annu Rev Pharmacol Toxicol 1995;35:581–606. Perlini S, Khoury EP, Norton GR, Chung ES, Fenton RA, Dobson Jr JG, Meyer TE. Adenosine mediates sustained adrenergic desensitization in the rat heart via activation of protein kinase C. Circ Res 1998;83:761– 771. Poon A, Sawynok J. Antinociception by adenosine analogs and inhibitors of adenosine metabolism in an inflammatory thermal hyperalgesia model in the rat. Pain 1998;74:235–245. Priebe TS, Nelson JA. Adenosine and immune system function. In: Phillis JW, editor. Adenosine and adenine nucleotides as regulators of cellular function, Boca Raton, FL: CRC Press, 1991. pp. 141–154. Ramsay MAE, Savege TM, Simpson BRJ, Goodwin R. Controlled sedatioin with alphaxalone–alphadolone. Br Med J 1974;2:656–659.

Reeve AJ, Dickenson AH. The roles of spinal adenosine receptors in the control of acute and more persistent nociceptive responses of dorsal horn neurones in the anaesthetized rat. Br J Pharmacol 1995;116:2221–2228. Sajjadi FG, Takabayashi K, Foster AC, Domingo RC, Firestein GS. Inhibition of TNF-a expression by adenosine. Role of A3 adenosine receptors. J Immunol 1996;156:3435–3442. Salter MW, De Koninck Y, Henry JL. An inhibitory postsynaptic potential in spinal nociceptive neurones is mediated by adenosine through activation of ATP-sensitive K 1 channels. Drug Dev Res 1993;28:416–422. Sawynok J. Adenosine receptor activation and nociception. Eur J Pharmacol 1998;317:1–11. Schraag S, Checketts MR, Kenny GNC. Lack of rapid development of opioid tolerance during alfentanil and remifentanil infusions for postoperative pain. Anesth Analg 1999;89:753–757. Segerdahl M, Ekblom A, Sandelin K, Wickman M, Sollevi A. Peroperative adenosine infusion reduces the requirements for isoflurane and postoperative analgesics. Anesth Analg 1995;80:1145–1149. Segerdahl M, Persson E, Ekblom A, Sollevi A. Peroperative adenosine infusion reduces isoflurane concentrations during general anesthesia for shoulder surgery. Acta Anaesthesiol Scand 1996;40:792–797. Segerdahl M, Irestedt L, Sollevi A. Antinociceptive effect of perioperative adenosine infusion in abdominal hysterectomy. Acta Anaesthesiol Scand 1997;41:473–479. Smith MA, Maehara Y, Ide Y, Sumida T, Collins JG, Kitahata LM. Suppression of rat spinal WDR neuronal responses to visceral stimulation by the adenosine receptor agonist l-phenyl-isopropyl adenosine. Anesth Analg 1996;82:S420. Sullivan GW, Linden J. Role of A2A adenosine receptors in inflammation. Drug Dev Res 1998;45:103–112. Sumida T, Smith MA, Maehara Y, Collins JG, Kitahata LM. Spinal Rphenyl-isopropyl adenosine inhibits spinal dorsal horn neurons responding to noxious heat stimulation in the absence and presence of sensitization. Pain 1998;74:307–313. Thiel M, Chouker A. Acting via A2 receptors, adenosine inhibits the production of tumor necrosis factor-a of endotoxin-stimulated human polymorphonuclear leukocytes. J Lab Clin Med 1995;126:275–282. Utting JE, Smith JM. Postoperative analgesia. Anaesthesia 1979;34:320– 332. Vinik HR, Kissin I. Rapid development of tolerance to analgesia during remifentanil infusion in humans. Anesth Analg 1998;86:1307–1311. Williams M. Adenosine: the prototypic neuromodulator. Neurochem Int 1989;14:249–264. Woolf CJ, Chong MS. Preemptive analgesia – treating postoperative pain by preventing the establishment of central sensitization. Anesth Analg 1993;77:362–379. Za´ rate E, Sa´ Re´ go MM, White PF, Duffy L, Shearer VE, Griffin JD, Whitten CW. Comparison of adenosine and remifentanil infusions as adjuvants to desflurane anesthesia. Anesthesiology 1999;90:956–963.