Intraoperative naloxone reduces remifentanil-induced postoperative hyperalgesia but not pain: a randomized controlled trial

Abstract

Background

Intraoperative use of a high-dose remifentanil may induce postoperative hyperalgesia. Low-dose naloxone can selectively reverse some adverse effects of opioids without compromising analgesia. We thus hypothesized that the intraoperative use of a high-dose remifentanil combined with a low-dose naloxone infusion reduces postoperative hyperalgesia compared with the use of remifentanil alone.

Methods

Patients undergoing elective thyroid surgery were randomly assigned into one of three groups, depending on the intraoperative effect-site concentration of remifentanil, with or without a continuous infusion of naloxone: 4 ng ml⁻¹ remifentanil with 0.05 μg kg⁻¹ h⁻¹ naloxone in the high-remifentanil with naloxone group, and 4 or 1 ng ml⁻¹ remifentanil with a placebo in the high- or low-remifentanil groups, respectively. We measured the pain thresholds (primary outcome) to mechanical stimuli using von Frey filaments and incidence of hyperalgesia on the peri-incisional area 24 h after surgery. We also measured pain intensity, analgesic consumptions and adverse events up to 48 h after surgery.

Results

The pain threshold presented as von Frey numbers [median (interquartile range)] was significantly lower in the high-remifentanil group (n=31) than in the high-remifentanil with naloxone (n=30) and the low-remifentanil (n=30) groups [3.63 (3.22–3.84) vs 3.84 (3.76–4.00) vs 3.80 (3.69–4.08), P=0.011]. The incidence of hyperalgesia was also higher in the high-remifentanil group than in the other groups [21/31 vs 10/30 vs 9/30, P=0.005]. Postoperative pain intensity, analgesic consumptions and adverse events were similar between groups.

Conclusions

The intraoperative use of low-dose naloxone combined with high-dose remifentanil reduced postoperative hyperalgesia but not pain.

Clinical trial registration

NCT02856087.

Editor’s key points

• Remifentanil has been associated with postoperative hyperalgesia.

• Von Frey filaments are used to measure mechanical nociceptive thresholds.

• Hyperalgesia is but one component influencing postoperative pain intensity.

• Low-dose naloxone can reduce remifentanil-induced hyperalgesia.

Remifentanil is a μ-opioid receptor agonist, which is widely used for intraoperative analgesia.^1–5 Remifentanil has unique pharmacokinetic properties with rapid onset and offset, thus higher doses can also be safely used without delaying postoperative recovery.^2,4,6,7 However, exposure to high doses of remifentanil may paradoxically reduce the pain threshold after its discontinuation.^{1,2,4–6,8–10} The postoperative hyperalgesia is known to be associated with acute and persistent pain.^11–13

Naloxone is a μ-opioid receptor antagonist commonly used to reverse several side-effects of opioids.^14,15 Low doses of naloxone can selectively eliminate adverse effects of opioids without compromising analgesia.^1,8,16–18 However, the effects of naloxone on remifentanil-induced hyperalgesia have not yet been investigated in surgical patients although its effects were suggested in an animal study.⁵ Therefore, we conducted a randomized trial to test the hypothesis that the intraoperative use of a high-dose remifentanil combined with a low-dose naloxone reduces postoperative hyperalgesia compared with the use of remifentanil alone.

Methods

Design

This prospective, double-blind, single-centre, parallel-group, randomized controlled trial was approved by the Institutional Review Board of Seoul National University Hospital (Seoul, Korea) and registered at ClinicalTrials.gov (NCT02856087). After obtaining written informed consent, we enrolled patients with ASA physical status I–II, aged 20–75 yr, and undergoing elective thyroid surgery under general anaesthesia using desflurane and remifentanil from November 2014 to April 2015. We excluded patients with an allergy to anaesthetic drugs, a history of thyroid surgery, current analgesic medications, drug or alcohol dependence, neurological or psychiatric disorders, severe hepatic or renal dysfunction, pregnancy and a BMI of >30 kg m⁻². Patients were randomly assigned in a 1:1:1 ratio with a computer-generated random sequence and the sealed envelope method by an assistant not involved in the trial. All investigators and patients were blinded to the group assignment.

Anaesthesia

Nurses not involved in the study prepared remifentanil with naloxone or normal saline according to the group assignment, which was blinded to the investigators. Two investigators (C.-H.K. and B.-R.K.) performed anaesthetic management according to a predetermined protocol. Without premedication, patients were monitored with non-invasive blood pressure, pulse oximetry, electrocardiography, bispectral index (A-2000 XP; Aspect Medical Systems, Newton, MA, USA) and acceleromyography (TOF-watch SX; MSD, Haarlem, the Netherlands), and received forced-air warming. Remifentanil (Ultiva; GlaxoSmithKline, Brentford, Middlesex, UK) was administered intravenously via effect-site target-controlled infusion (Base Primea; Fresenius Vial, Brézins, France) with Minto pharmacokinetic model, and naloxone (Samjin Pharm, Seoul, Korea) or normal saline were given at a constant infusion rate according to the group assignment. Patients were randomly assigned to one of three groups depending on the effect-site concentration of remifentanil with or without a continuous infusion of naloxone during anaesthesia: 4 ng ml⁻¹ of remifentanil with 0.05 μg kg⁻¹ h⁻¹ of naloxone in the high-remifentanil with naloxone group; and 4 or 1 ng ml⁻¹ of remifentanil with placebo in the high- and low-remifentanil groups, respectively. The study drugs were administered from anaesthetic induction until skin closure after surgery. The infusion devices were blinded to the investigators.

General anaesthesia was induced by i.v. propofol 1.5–2 mg kg⁻¹ and rocuronium 0.6–0.8 mg kg⁻¹. At a bispectral index of <60 and train-of-four count=0, the patient’s trachea was intubated with a direct or video (UEscope; UE Medical Devices, Newton, MA, USA) laryngoscope. Reinforced tracheal tubes (Medtronic, Minneapolis, MN, USA) with an inner diameter of 7.0 and 7.5 mm were used for women and men, respectively. The intracuff pressure of the tube was adjusted to less than 25 cm H₂O with a cuff pressure monitor (VBM Medizintechnik GmbH, Sulz am Neckar, Germany). Core temperature was monitored in the nasopharynx.

Anaesthetic depth was maintained at a bispectral index of 40–60 by adjusting the end-tidal concentration of desflurane. The patient’s lungs were ventilated (Primus; Dragger, Lubeck, Germany) with a tidal volume 6–8 ml kg⁻¹ of predicted body weight, a positive end-expiratory pressure 5 cm H₂O, and an inspired oxygen fraction 0.5 with a fresh gas flow 2 litres min⁻¹ of oxygen and air. The respiratory rate was set to obtain an end-tidal carbon dioxide tension of 4.7–5.3 kPa. Rocuronium 0.2–0.3 mg kg⁻¹ was intermittently administered at a train-of-four count ≥1. At a mean blood pressure of <60 mm Hg, lactated Ringer’s solution 200 ml, ephedrine 5 mg or phenylephrine 30 μg were given as appropriate.

At skin closure after surgery, administration of desflurane, remifentanil and naloxone or placebo was stopped, and ramosetron 0.3 mg and ketorolac 30 mg were given intravenously. At a train-of-four ratio of >90%, pyridostigmine 300 μg kg⁻¹ and glycopyrrolate 10 μg kg⁻¹ were administered to reverse the residual neuromuscular block. Confirming adequate spontaneous breathing and responses to verbal commands, the patient was transferred to the post-anaesthesia care unit after tracheal extubation. The patient was discharged from the unit at a modified Aldrete score of 9 or 10.

Outcomes

We collected baseline data on patients, surgery and anaesthesia. We checked amounts of anaesthetic drugs or fluid and requirements for inotropes during surgery. We also recorded mean blood pressure, heart rate and end-tidal concentrations of desflurane at six time points: before induction and intubation, one minute after intubation, skin incision, skin closure, and extubation. An investigator (S.Y.) evaluated pain intensity using an 11-point numeric rating scale (0, no pain; 10, worst pain imaginable) at six time points: 30 min, 1, 6, 12, 24, and 48 h after surgery. Patients with the pain score higher than 4 received i.v. or i.m. ketorolac 30 mg and the number of rescue analgesic medications was recorded. We checked lengths of stay in the post-anaesthesia care unit and hospital.

The investigator (S.Y.) assessed pain thresholds to mechanical stimuli on the peri-incisional area of thyroid surgery and the forearm preoperatively, at 24 and 48 h after surgery as described in previous studies.^4,19,20 The pain threshold was measured with von Frey filaments (Touch Test^TM Sensory Evaluators; Stoelting Co., Wood Dale, IL, USA) and presented as von Frey numbers logarithmically transformed from the mechanical force produced by the filament.^21–23 The lower von Frey number indicates the lower threshold and more sensitization to pain.^21–23 On the peri-incisional area, the pain threshold was defined as the average of three measurements at 2 cm below the bilateral edges and middle of the skin incision. The incision length was 5–6 cm in all patients. The threshold was also obtained on the non-dominant forearm by averaging three measurements at 3, 6 and 9 cm distal to the antecubital crease. We checked numbers of patients with lower, higher and same postoperative pain thresholds compared with the preoperative values on the peri-incisional area and forearm. Postoperative hyperalgesia was defined as a decrease in the postoperative pain threshold compared with the preoperative one.

The primary outcome was the pain threshold on the peri-incisional area 24 h after surgery. Secondary outcomes were the other pain thresholds and incidence of hyperalgesia on the peri-incisional area or forearm; postoperative rescue analgesic requirements, pain intensity and adverse events; and intraoperative haemodynamic variables and amounts of drugs or fluid.

Statistical analysis

In our pilot study (n=10), the mean (sd) von Frey number on the peri-incisional area was 3.1 (0.8) at 24 h after thyroid surgery using desflurane and 4 ng ml⁻¹ effect-site concentration of remifentanil. For a clinically significant 20% difference in the pain threshold by adding naloxone, 27 patients were needed in each group with a risk of type-I error of 0.05 and power of 0.8 for two-tailed statistical analysis.

Continuous variables were presented as mean (sd) or median (interquartile range) after checking the normality with the Shapiro–Wilk test. Outcomes collected sequentially were analysed with repeated-measures analysis of variance (ANOVA) or the Friedman test for entire time points, and then with one-way ANOVA or Kruskal–Wallis test, unpaired or paired t-tests, and Mann–Whitney U or Wilcoxon signed-rank tests at each time point as appropriate. Categorical variables were the number of patients compared with Fisher’s exact test. Effect sizes with 95% confidence interval (CI) were calculated if necessary. All analyses were conducted in an intention-to-treat manner. A P-value <0.05 was considered statistically significant. STATA software (Stata Corporation, College Station, TX, USA) was used for all statistical analyses, sample size calculation and randomization.

Results

After screening 95 patients, 91 eligible patients were included, 30 in the high-remifentanil with naloxone group, 31 in the high-remifentanil group and 30 in the low-remifentanil group (Fig. 1). Characteristics of patients, surgery and anaesthesia were comparable between groups except doses of remifentanil and naloxone (Table 1). The end-tidal concentrations of desflurane were similar during anaesthesia (P=0.28 by repeated-measures ANOVA). However, the mean blood pressure (Fig. 2A;P=0.001 by repeated-measures ANOVA) and heart rate (Fig. 2B;P=0.004) at tracheal intubation and skin incision were significantly higher in the low-remifentanil group than in the high-remifentanil with and without naloxone groups: 108 (31) vs 88 (25) vs 89 (25) mm Hg, P=0.007 and 91 (21) vs 81 (18) vs 76 (18) beats min⁻¹, P=0.008 at intubation; 91 (21) vs 76 (17) vs 74 (14) mm Hg, P<0.001 and 78 (18) vs 65 (11) vs 65 (12) beats min⁻¹, P<0.001 at incision by one-way ANOVA.

Table 1

Characteristic of patients, surgery and anaesthesia. Data are number of patients or mean (sd) except age [mean (range)]. *Significantly lower in the low-remifentanil group than in the other groups by one-way analysis of variance (ANOVA) and unpaired t-test. ^†Only given in the high-remifentanil with naloxone group

	High-remifentanil with naloxone group (n=30)	High-remifentanil group (n=31)	Low-remifentanil group (n=30)	P-value
Age (yr)	50 (24–71)	44 (24–75)	52 (22–75)	–
Sex (male/female)	8/22	12/19	9/21	–
Weight (kg)	63 (11)	63 (13)	61 (11)	–
Height (cm)	160 (8)	165 (10)	161 (8)	–
BMI (kg m–2)	24.4 (3.4)	23.1 (3.2)	23.4 (2.9)	–
ASA physical status (I/II)	25/5	25/6	20/10	–
Medical conditions (hypertension/diabetes/pulmonary disease)	4/3/3	6/1/5	8/2/3	–
Type of surgery (thyroidectomy/lobectomy)	25/5	21/10	26/4	–
Amount of anaesthetic drugs and a fluid
Propofol (mg)	105 (28)	100 (21)	114 (23)	0.095
Rocuronium (mg)	49 (9)	49 (11)	47 (7)	0.64
Remifentanil (μg)*	1029 (334)	1101 (451)	321 (202)	<0.001
Naloxone (μg)†	5.5±2.4	0	0	<0.001
Lactated Ringer’s solution (ml)	417 (167)	405 (165)	455 (259)	0.60
Patients receiving inotropes	19	22	19	0.76
Duration of surgery (min)	92 (21)	89 (41)	94 (37)	0.86
Duration of anaesthesia (min)	122 (24)	124 (42)	124 (41)	0.96

Fig 1

CONSORT diagram.

Fig 2

Mean blood pressure (A) and heart rate (B) during anaesthesia. Data are mean and sd. *Mean difference (95% CI) 20 (5–35) mm Hg, P=0.008 and 19 (5–34) mm Hg, P=0.010 compared with the high-remifentanil with and without naloxone groups, respectively, by unpaired t-test; ^†14 (5–24) mm Hg, P=0.005 and 17 (8–26) mm Hg, P<0.001; ^‡10 (1–20) beats min⁻¹, P=0.043 and 15 (5–25) beats min⁻¹, P=0.003; ^§14 (5–24) beats min⁻¹, P=0.001 and 14 (4–26) beats min⁻¹, P=0.002.

On the peri-incisional area, the pain thresholds presented as von Frey numbers (Fig. 3A;P=0.046 by Friedman test) were significantly lower in the high-remifentanil group than in the high-remifentanil with naloxone and the low-remifentanil groups at 24 and 48 h after surgery: 3.63 (3.22–3.84) vs 3.84 (3.76–4.00) vs 3.80 (3.69–4.08), P=0.011 at 24 h; 3.61 (3.22–3.84) vs 3.84 (3.76–4.00) vs 3.90 (3.69–4.08), P=0.004 at 48 h by Kruskal–Wallis test. In the pairwise comparisons within each group, the postoperative pain thresholds were significantly lower than the preoperative one in the high-remifentanil group (P=0.006 at 24 h; P=0.005 at 48 h by Wilcoxon signed-rank test), but comparable in the other groups. The incidence of postoperative hyperalgesia on the peri-incisional area was also significantly higher in the high-remifentanil group than in the other two groups (Table 2). However, the postoperative pain threshold (Fig. 3B;P=0.345 by Friedman test) and the incidence of hyperalgesia (Table 2) on the forearm were comparable between and within groups. In addition, there were no significant differences in the postoperative pain intensity expressed as numeric rating scales (Fig. 4), rescue analgesic requirements, adverse events and lengths of stay in the post-anaesthesia care unit or hospital (Table 2).

Table 2

Postoperative outcomes. Data are number of patients or mean (sd). *Numbers of patients with lower, higher and same postoperative pain thresholds compared with the preoperative values. ^†P=0.016 and 0.007 and ^‡P=0.001 and <0.001 compared with the high-remifentanil with naloxone and the low-remifentanil groups, respectively, by Fisher’s exact test

	High-remifentanil with naloxone group (n=30)	High-remifentanil group (n=31)	Low-remifentanil group (n=30)	P-value
Patient receiving analgesics	9	13	8	0.46
Pain threshold on the peri-incisional area* (lower/higher/same)
24 h after surgery	10/7/13	21/5/5†	9/6/15	0.018
48 h after surgery	10/5/15	23/5/3‡	8/7/15	0.001
Pain threshold on the forearm* (lower/higher/same)
24 h after surgery	7/3/20	10/6/15	6/6/18	0.56
48 h after surgery	7/5/18	11/5/15	6/4/20	0.63
Adverse events
Nausea	5	5	2	0.47
Dizziness	5	3	9	0.13
Headache	5	6	5	>0.99
Drowsiness	2	2	2	>0.99
Shivering	2	3	0	0.36
Length of stay
Post-anaesthetic care unit (min)	44 (6)	46 (7)	45 (3)	0.38
Hospital (days)	4.1 (0.7)	4.0 (0.7)	4.3 (0.9)	0.40

Fig 3

Pain thresholds to mechanical stimuli presented as von Frey numbers on the peri-incisional area of thyroid surgery (A) and the forearm (B). P=0.007 (*), 0.016 (†) and 0.004 (‡) by Mann–Whitney U-test.

Fig 4

Postoperative pain intensity assessed by an 11-point numeric rating scale (0, no pain; 10, worst pain imaginable). Data are mean and sd. There were no significant differences between groups [P=0.23 by repeated-measures analysis of variance (ANOVA)].

Discussion

Remifentanil provides sufficient intraoperative analgesia attenuating haemodynamic fluctuations during surgery.^1–3,6 Because of its short context-sensitive half-time of less than 9 min,⁷ high doses of remifentanil can be safely given intraoperatively with little risk of delayed recovery or adverse effects postoperatively.^2,4,6,7 In our study, the intraoperative use of the high-dose remifentanil (4 ng ml⁻¹ of effect-site concentration) compared with the low-dose (1 ng ml⁻¹) attenuated increases in the mean blood pressure and heart rate induced by noxious stimuli such as tracheal intubation or skin incision. The requirements for inotropes or the fluid, length of stay in the post-anaesthesia care unit or hospital, and adverse events were similar regardless of the intraoperative dose of remifentanil. Therefore, the 4 ng ml⁻¹ effect-site concentration of remifentanil seems to be effective for adequate intraoperative analgesia in relatively healthy patients, although caution should be taken in cardiovascular-compromised patients.^24,25

However, the intraoperative use of remifentanil can lead to postoperative hyperalgesia.^2,4,6 The hyperalgesia is known to be induced by an effect-site concentration of remifentanil higher than 2.7 ng ml⁻¹,² and its underlying mechanisms are likely to be associated with upregulation or alteration of N-methyl-D-aspartate (NMDA)^8,26–28 or μ-opioid receptors.^5,29 In our study, the intraoperative use of the high-dose remifentanil compared with the low dose decreased the pain thresholds and increased the incidence of hyperalgesia on the peri-incisional area, in agreement with previous studies.^30,31

Our study showed that adding the low-dose naloxone (0.05 μg kg⁻¹ h⁻¹) to the high-dose remifentanil during anaesthesia significantly reduced postoperative hyperalgesia on the peri-incisional area compared with using remifentanil alone. This anti-hyperalgesic effect of naloxone is likely to be produced by antagonizing or modifying NMDA^4,8,26,32 and μ-opioid receptor activities^5,17,33,34 related to the opioid-induced hyperalgesia, although our study did not investigate its precise mechanisms. Naloxone can reverse the analgesic effects of opioids as well as the side-effects in doses higher than 0.06 μg kg⁻¹ h⁻¹,¹⁷ thus lower doses may be preferable to selectively obtain anti-hyperalgesic effects without affecting analgesia.

The intraoperative opioids and nociceptive stimuli may have synergistic effects on postoperative hyperalgesia.^4,13,23 This might explain no difference in the pain threshold on the forearm unrelated to surgical insult in our study, in accordance with previous findings.^4,23,35,36 Furthermore, hyperalgesia is known to be more associated with pain evoked by coughing or movements rather than resting pain.^4,37 In addition, although hyperalgesia may aggravate postoperative pain,^6,10 the association was reported to be low between the pain threshold objectively measured by von Frey filaments and the pain intensity subjectively assessed by the Likert-type scale.^19,31,38 They might explain the similar pain intensity presented as the numeric rating scale despite the difference in the pain threshold.^19,31,38

Our study has some limitations. As mentioned above, we only checked the pain intensity in a resting state but not during coughing or movements. We also only collected outcomes for intraoperative and acute postoperative periods although hyperalgesia is known to be associated with chronic pain.^12,13 Therefore, further research is needed to investigate the effects of intraoperative naloxone on acute and chronic postoperative pain.

In conclusion, the intraoperative use of low-dose naloxone combined with high-dose remifentanil reduced postoperative hyperalgesia compared with the use of remifentanil alone, although it seemed to have no clinical benefits in reducing postoperative pain.

Authors’ contributions

Study design: C.-H.K, S.Y., T.K.K., Y.J., J.-H.S.

Study conduct and data collection: C.-H.K., S.Y., B.-R.K.

Data analysis: Y.J.C., T.K.K., Y.J.

Writing paper: C.-H.K., S.Y., Y.J.C., J.-H.S.

Revising paper: all authors

Acknowledgements

We thank the American alumni association of our department for the support of this study.

Declaration of interest

None declared.

Funding

This work was supported by the Department of Anaesthesiology and Pain Medicine, Seoul National University Hospital.

Author notes

†

These two authors equally contributed to this work as co-first authors.