Secondary Logo

Journal Logo


Assessing the antinociceptive effect of nitrous oxide to tetanic stimulation in anaesthetised patients with new intra-operative nociception monitors

An observational study

Coulombe, Marie-Andrée; Décary, Elizabeth; Maximos, Sarah; Brulotte, Véronique; Drolet, Pierre; Tanoubi, Issam; Issa, Rami; Zaphiratos, Valérie; Verdonck, Olivier; Fortier, Louis-Philippe; Godin, Nadia; Idrissi, Moulay; Raft, Julien; Richebé, Philippe

Author Information
European Journal of Anaesthesiology: May 2021 - Volume 38 - Issue 5 - p 512-523
doi: 10.1097/EJA.0000000000001431
  • Free



Nitrous oxide (N2O) has been used in anaesthesia since the 19th century. It is known for its anxiolytic, analgesic, euphoric and hypnotic properties in awake patients. N2O is often co-administered with other hypnotic agents and augments the concentrations of other drugs required for induction and maintenance of anaesthesia by nearly 35%.1 It has also been reported that there is a meaningful effect of N2O on the transmission of noxious stimuli through its action on descending pain modulation systems leading to endogenous opioid peptide release in the peri-aqueductal grey matter of the midbrain,2–4 and on multiple ion channel conductances, acting predominantly as a N-methyl-d-aspartate (NMDA) receptor antagonist.5–7 As a direct consequence, N2O is able to oppose the development of postoperative pain sensitisation,8 reduce opioid-induced hyperalgesia9–13 and may even reduce the incidence of persistent postsurgical pain.14–16

However, no recent study has reported comprehensive quantification of the antinociceptive effect of N2O during general anaesthesia, more commonly known by anaesthesia providers as an ‘analgesic’ effect. This was mostly due to a lack of efficient intra-operative nociception monitoring. The term intra-operative nociception refers to noxious stimuli encoded and processed by the central nervous system under general anaesthesia.17 To date, responses to noxious stimuli during surgery were commonly assessed using blood pressure (BP) and heart rate (HR), or indirectly through anaesthetic and analgesic consumption during anaesthesia. Unsurprisingly, those criteria, with low sensitivity and specificity to detect noxious stimulus under general anaesthesia,18–21 were inadequate in addressing the question of whether N2O had a significant clinical analgesic effect during surgery.

Recently, two new indices have been developed to monitor intra-operative pain in patients under general anaesthesia such as the analgesia nociception index (ANI) and the nociception level (NOL) index. The ANI is based on a single parameter analysis: the HR variability (HRV), the analysis of R–R series variability from the ECG, which relates to the activity of the autonomic nervous system (ANS). The ANI evaluates the ANS tone on a numerical scale ranging between 0 and 100 (100 being a total absence of nociception and 0 maximal nociception), and has been shown to correlate with pain/nociception-analgesia in awake and anaesthetised subjects.22,23 The NOL is a multiparameter index incorporating HR, HRV, plethysmograph wave amplitude, skin conductance level, number of skin conductance fluctuations and their time derivatives through a nonlinear combination. These are all obtained through a noninvasive finger probe. The NOL index is also displayed as a single number from 0 to 100 (absence to severe noxious stimulation, respectively). The NOL index correlates well with noxious stimuli of various intensities in studies of patients undergoing elective surgery under general anaesthesia. In addition, the NOL performed better than the classically used clinical parameters18–21 such as HR and BP.

Therefore, the aim of this study was to assess the antinociceptive effect of N2O quantitatively in a clinical setting with patients undergoing laparotomy for general surgery under general anaesthesia by using a combination of both classical parameters and the new indices, ANI and NOL.


The study was a prospective, open label, observational and descriptive study, approved by the local Scientific and Ethic Committee [Comité d’Éthique de la Recherche, Hôpital Maisonneuve-Rosemont, CIUSSS de l’Est de l’île de Montréal (CEMTL), Montréal, QC, Canada; CER 15064, 9 December 2015] and was registered under the number NCT02701478 at All patients were recruited at the University Hospital of Maisonneuve-Rosemont, CEMTL, Université de Montreal, Montreal, Canada, between January and September 2016, and they all gave their written informed consent.

The study included 40 patients, 19 men and 21 women, having elective abdominal surgery with laparotomy under epidural analgesia and general anaesthesia. To be included, patients had to have an American Society of Anesthesiologists (ASA) physical status of 1, 2 or 3, and were free from any condition where internal entrapped air could expand and become dangerous. Patients were not included if they had a history of coronary artery disease, any cardiac arrhythmia (including atrial fibrillation), chronic use of psychotropic and/or opioid drugs, use of drugs that affect the ANS (including β-blockers), allergies to any drug used in the study protocol, or had received intra-ocular injection of gas (such as SF6, C3F8, C2F6) within 3 months before surgery. Moreover, patients were excluded if 100% oxygen (O2) ventilation was required during anaesthesia (except for the induction phase during which all patients received inhaled 100% O2), if the patient requested unexpected supplemental airway manipulations due to physical complications, or if strong haemodynamic support was required following unexpected complications (such as massive blood loss and transfusions, volume challenges, vasopressors, major inotropic drug support).

The study procedures are summarised in Fig. 1. On arrival in the operating room, and before induction of general anaesthesia, an intravenous cannula was inserted and a lower thoracic epidural (between T9 and T12) was performed with the patient in the sitting position. Intravenous midazolam 1 mg was administered at the time of epidural insertion, but no intravenous opioids were administered. Subsequently, monitoring comprising noninvasive BP, ECG and SpO2 was started. Intravascular positioning of the epidural catheter was excluded by gentle suction with 1-ml syringe and followed by a 3-ml test dose of lidocaine 2% with epinephrine 1/200 000 (AstraZeneca Inc, Longueuil, Quebec, Canada). Patients were then equipped with both nociception monitoring devices – the PhysioDoloris monitor (Metrodoloris Medical Systems, Lille, France) and the PMD-200 monitor (Medasense Biometrics Ltd, Ramat Gan, Israel) – which measure the ANI and NOL indices, respectively. Neither paracetamol nor NSAIDs were given prior to surgery as all patients benefited intra-operatively from epidural analgesia. Both paracetamol and NSAIDs were started at the end of surgery if no contraindications were present.

Fig. 1:
Study design and data collection timeline

General anaesthesia was induced with propofol 1.5 to 2.5 mg kg−1 (Pharmascience Inc, Montreal, Quebec, Canada) and remifentanil 1 μg kg−1 as an intravenous bolus, then at an infusion rate of 0.05 μg kg−1 min−1 (Teva, Toronto, Ontario, Canada), and rocuronium 0.6 to 1 mg kg−1 (Sandoz Canada Inc, Boucherville, Quebec, Canada) was given to produce muscle relaxation. Tracheal intubation was performed once no response to train-of-four ulnar nerve stimulation was detected. Once intubated, the remifentanil infusion was decreased to a very low dose (0.005 μg kg−1 min−1) and the patient was not manipulated by any member of the operating room team for the following 5 min (assessment of postintubation study parameters). General anaesthesia was maintained using inhaled desflurane and the end-tidal fraction (end-tidal desflurane) was adjusted to maintain a bispectral index (BIS) value between 45 and 55. As seen in Fig. 1, once intubated, and after recording all the study parameters for 5 min, the epidural space was loaded through the epidural catheter with a total of 8 to 10 ml of lidocaine 2% and 1/200 000 epinephrine (AstraZeneca Inc) over a period of 10 min in preparation for the first incision.

The electrical stimuli were delivered by a nerve stimulator applied on the forearm using a standardised electrical current [tetanic stimulation at 70 mA, 100 Hz for 30 s (EZstim II; Life-Tech, Stafford, Texas, USA)].20,21,24 All electrical tetanic stimulations were performed after the first incision (Fig. 1) without any other manipulation of the patient (the surgery was temporarily paused) for at least 2 min prior to each tetanic stimulation and for 3 min after each stimulation. The remifentanil infusion was set to a very low dose (0.005 μg kg−1 min−1) at least 10 min before each electrical tetanic stimulation. This practice provided for a minimal impact on the poststimulation response (see our previous study20,21). HR variation, mean arterial BP (MAP), BIS index, ANI and NOL indices were measured for 2 min preceding and 3 min following each stimulation. This procedure was carried out under three conditions of inhaled gas: 0% N2O/50% O2 (Stimulation 1), 50% N2O/50% O2 (Stimulation 2) and 25% N2O/50% O2 (Stimulation 3). At our institution, gases (O2, air, N2O) were provided by Praxair (Praxair Canada Inc., Québec City, Québec, Canada) and were administered through the breathing circuit of a Dräger Ventilator (Perseus A500, Dräger, Germany). Stimulation sequences were the same for all patients. For each stimulation, end-tidal-N2O was stabilised for 5 min at the appropriate concentration (0, 25 or 50%) before any stimulus was applied and any measurement was taken (Table 1). At the end of the protocol, and once the three periods of stimulations were completed, the anaesthesia management was left at the discretion of the anaesthesiologist in charge of the patient for the rest of the operation and data collection ended.

Table 1 - Gas analysis at the time of each tetanic electrical stimulation
Stim. 1: 0% N2O in 50% O2 Stim. 2: 50% N2O in O2 Stim. 3: 25% N2O in O2
Inspired O2 (%) 50.8 [46.7 to 58.0]: 0 vs. 50%, P = 0.0017∗∗: 0 vs. 25%, P = 0.0017∗∗∗: 25 vs. 50%, P < 0.0001 37.7 [34.4 to 40.1] 63.2 [62.1 to 65.3]
End-tidal desflurane (%) 4.2 [3.8 to 5.1]: 0 vs. 50%, P < 0.0001∗∗: 0 vs. 25%, P = 0.0108∗∗∗: 25 vs. 50%, P = 0.0108 2.8 [2.4 to 3.4] 3.5 [2.9 to 4.5]
MAC desflurane (%) 0.75 [0.70 to 0.90]: 0 vs. 50%, P < 0.0001∗∗: 0 vs. 25%, P = 0.0007∗∗∗: 25 vs. 50%, P = 0.0026 1.00 [0.93 to 1.16] 0.90 [0.80 to 1.00]
End-tidal N2O (%) 0.00 [0.00 to 0.00]: 0 vs. 50%, P < 0.0001∗∗: 0 vs. 25%, P < 0.001∗∗∗: 25 vs. 50%, P < 0.0001 51.1 [49.0 to 52.7] 26.0 [25.8 to 27.4]
Values are median [IQR]. MAC, minimum alveolar concentration; Stim., tetanic electrical stimulation of the forearm.

All intra-operative data were recorded electronically by an automated information management system. Every 5 s, intra-operative NOL and ANI values were recorded. Other intra-operative parameters evaluated in this study – HR, MAP, BIS index, respiratory parameters and volatile anaesthetics concentrations – were automatically and electronically collected every second from a Perseus A500 anaesthesia workstation combined to the physiological monitor M540 (Dräger AG & Co., Lübeck, Germany). At the end of the procedure, all data were anonymously transferred and stored into a research computer data file for statistical analysis.

Statistical analysis

For this study, our main hypothesis was that following the nociceptive stimulus, the variation of the ANI index (ΔANI) would be significantly lower when N2O was administered to the patients at 50% inhaled concentration compared with when no N2O was given.

ANI index decreases when nociception occurs whereas NOL index, HR and mean BP all increase. Our preliminary results showed that the mean ± SD decrease of ANI when N2O was not administered (0% N2O/50% O2) was 30 ± 15 arbitrary units after the electrical tetanic stimulation. When end-tidal-N2O was set at a concentration of 50% (combined with 50% O2), ΔANI was expected to be only 15 ± 15, 50% less reactive to the same nociceptive stimulation administered to an anaesthetised patient.

Our preliminary results illustrated that, with a sample size calculation, 0.05 type I (α) error, a power (1-β) of 0.80 and a bilateral test, 32 patients would be needed to reach our primary objective. Forty patients were included in the study to allow for potential withdrawals.

As previously mentioned, data from the NOL and ANI were collected every 5 s, and the last minute of the baseline (prior to nociceptive clinical or experimental stimulation) was averaged. To pair the NOL and ANI data, for the BIS, HR and MAP, only paired values, recorded every 5 s were considered. Peak data were identified manually and confirmed using the automatic peak function of GraphPad Prism (GraphPad Prism version 8.01 for Windows; GraphPad Software, La Jolla, California, USA) for each parameter. Data from the peak value ±10 s (two data points) before and after the peak were averaged (five data points total) to avoid a peak due to any artefact for each of the recorded parameters. Absolute differences (Δ for each parameter) were calculated using an average of the last minute of the baseline (before the stimulation) and the average of the five data points representing the peak poststimulation for each parameter.

All data related to respiratory parameters and volatile anaesthetics concentration were averaged 1 min before the stimulation and 3 min after the stimulation (Table 1).

Gaussian distribution of the data was evaluated using the Shapiro–Wilk normality test. Since parts of our data were not normally distributed, nonparametric tests were used. Pre vs. poststimulation comparisons for each parameter were analysed using Wilcoxon matched pairs signed rank test (nonparametric, paired analysis). Poststimulation responses to clinical stimuli (tracheal intubation and first incision) and experimental stimuli (tetanic electrical stimulation) were compared with their respective prestimulation measures also using Wilcoxon matched-pairs signed-rank test. To account for multiple comparisons, a Bonferroni correction (α/number of comparisons, α = 0.05) was applied. A total of five comparisons (two clinical and three experimental stimulations) were performed. Our results were considered significant if P was less than 0.01. The effect of N2O concentration on output fluctuations (NOL, ANI, HR, BIS and MAP, Fig. 2) were measured using a Friedman test (nonparametric, paired, ANOVA) followed by Dunn's post hoc test to determine significant differences between N2O concentrations. Statistical analyses were performed using IBM SPSS Statistics for Mac version 24.0 (IBM Corp., Armonk, New York, USA).

Fig. 2:
Evolution of each study parameter over time from 1 min before standardised tetanic electrical stimulation to 3 min after stimulation


As shown in Fig. 3, 49 patients were screened and data from 40 patients (19 men and 21 women) were analysed. The mean ± SD age was 54 ± 11 years old, mean height was 167.6 ± 9.6 cm, mean real weight, adjusted body weight and BMI were 74.3 ± 16.8, 65.9 ± 13.0 kg and 26.5 ± 5.9 kg m−2, respectively, and ASA status was ASA 1 for seven patients, ASA 2 for 34 patients and ASA 3 for five patients. Intravenous doses of the drugs used for induction of general anaesthesia were as follows: midazolam 1 mg to all patients at the epidural insertion, propofol 141 ± 32 mg, remifentanil 70 ± 18 μg, rocuronium 50 ± 8 mg. Epidural total dose of lidocaine 2% was given 3 ml immediately after epidural needle insertion, and the total lidocaine 2% dose administered before incision was 8 ± 0.9 ml. Most of the participants underwent midline laparotomy for bowel surgery (81%) with one-third undergoing tumour resection and hyperthermic intraperitoneal chemotherapy (HIPEC). For HIPEC procedures, all the study procedures listed in the protocol were performed before the intraperitoneal chemotherapy was started (e.g. during the cytoreduction phase). Other participants underwent gynaecological and urological procedures with a vertical abdominal midline incision.

Fig. 3:
Flow chart of the study

Table 1 shows the gas analysis for O2, desflurane and N2O at each experimental tetanic electrical stimulation (Stimulations 1 to 3). It illustrates that when administering N2O in the gas mixture given to the patient, the fraction of inspired O2 was significantly different between all three assessed times in the same patient. Also, when N2O was administered, the end-tidal concentration of desflurane decreased significantly whereas the expired fraction of desflurane significantly increased (Table 1).

In this study, variations of NOL index, ANI index, HR, MAP and BIS index during the baseline and postclinical stimulations were measured. The main difference between these two clinical stimulations was the loading of the epidural space with local anaesthetic after the intubation, but before the first incision (Fig. 1). Table 2 shows baseline values of parameters before and after epidural loading, both without any stimulation. The NOL index did not differ before or after the epidural space loading (P = 0.5919). Also, neither HR nor BIS changed after the epidural space loading (Table 2). The ANI baseline values did change, however, before and after the epidural space loading, demonstrating the impact of epidural local anaesthesia administration on this index (P = 0.0014).

Table 2 - Comparison of baseline values of each variable before the clinical stimulations of intubation and incision
Preepidural Postepidural Absolute difference
Median [IQR] Median [IQR] Mean (95% CI) P
NOL 7.3 [2.1 to 12.7] 4.0 [2.8 to 9.5] −0.08 (−5.7 to 5.5) 0.5919
ANI 50.8 [41.0 to 60.4] 71.3 [51.0 to 89.3] 14.9 (7.2 to 22.5) 0.0014
HR 65.9 [60.1 to 71.2] 62.4 [53.5 to 74.9] −1.7 (−5.9 to 2.6) 0.3276
BIS 46.7 [38.1 to 57.1] 42.8 [40.9 to 45.7] −3.4 (−9.3 to 2.5) 0.2522
Pre-epidural, baseline value for each variable before the clinical stimulus intubation; postepidural, baseline value for each variable before the clinical stimulus incision. ANI, analgesia nociception index; BIS, bispectral index; CI, confidence interval; HR, heart rate; NOL, nociception level index.

For the same five variables, we also measured the variation induced by two different clinical stimulations (intubation or first incision). Significant differences were observed for all variables for both intubation and incision (Table 3 and Fig. 4a).

Table 3 - Changes in each variable after different stimulus (intubation, incision and tetanic electrical stimulation) at various concentrations of inhaled N2O
Prestimulation Poststimulation Absolute difference
Type of stimulus Median [IQR] Median [IQR] Mean (95% CI) P
 Intubation 7.3 [2.1 to 12.7] 30.8 [18.2 to 49.1] 24.1 (17.2 to 31.1) <0.0001
 Incision 4.3 [2.8 to 9.5] 24.6 [14.0 to 37.2] 18.6 (13.3 to 23.9) <0.0001
 0% N2O 4.3 [3.0 to 7.9] 32.4 [22.6 to 46.6]A 26.6 (21.9 to 31.3)D <0.0001
 25% N2O 6.8 [2.8 to 9.6] 25.4 [15.2 to 35.8]B 18.5 (13.3 to 23.6)E <0.0001
 50% N2O 7.2 [3.2 to 12.8] 23.2 [16.2 to 36.8]C 18.4 (12.9 to 23.9)F <0.0001
P = 0.3012 (F = 1.48) P = 0.0009 (F = 14.13)B vs. A: 0.0018C vs. A: 0.0069 P = 0.0018 (F = 12.65)E vs. D: 0.0029F vs. D: 0.0156
 Intubation 46.8 [39.1 to 59.2] 35.2 [28.6 to 45.0] −11.1 (−15.1 to −7.0) <0.0001
 Incision 73.9 [52.7 to 87.7] 55.7 [45.8 to 72.1] −9.7 (−14.6 to −4.8) 0.001
 0% N2O 63.2 [50.8 to 70.5]G 36.6 [28.4 to 50.9]I −20.3 (−25.9 to −14.7) <0.0001
 25% N2O 77.0 [61.4 to 87.9]H 57.2 [46.6 to 76.1]J −13.7 (−20.2 to −7.1) 0.0001
 50% N2O 69.6 [59.7 to 84.0] 54.2 [44.3 to 66.0]K −17.0 (−24.0 to −10.1) <0.0001
P = 0.0060 (F = 10.22)H vs. G: 0.0076 P = 0.0001 (F = 23.5)J vs. I: <0.0001K vs. I: 0.0019 P = 0.1230 (F = 4.191)
 Intubation 65.8 [59.1 to 71.0] 76.6 [65.8 to 92.6] 13.9 (9.4 to 18.5) <0.0001
 Incision 62.4 [56.8 to 73.8] 64.1 [55.5 to 81.7] 5.2 (2.1 to 8.3) 0.0044
 0% N2O 66.8 [60.2 to 78.0]L 80.0 [69.6 to 88.2]O 11.3 (8.3 to 14.2)R <0.0001
 25% N2O 63.0 [56.3 to 72.7]M 72.5 [59.2 to 78.7]P 5.2 (3.5 to 6.9)S <0.0001
 50% N2O 69.4 [55.0 to 80.4]N 74.4 [62.9 to 83.6]Q 5.5 (3.3 to 7.7)T <0.0001
P = 0.0003 (F = 14.2)M vs. L: 0.0007N vs. L: 0.0027 P = 0.0003 (F = 16.05)P vs. O: 0.0008Q vs. O: 0.0032 P = 0.0003 (F = 15.94)S vs. R: 0.0021T vs. R: 0.0013
 Intubation 44.7 [38.0 to 56.8] 57.6 [46.0 to 68.3] 10.2 (6.8 to 13.6) <0.0001
 Incision 43.3 [40.9 to 45.7] 46.0 [42.6 to 51.0] 4.4 (2.5 to 6.3) <0.0001
 0% N2O 43.4 [41.1 to 45.6] 46.9 [43.8 to 54.8] 5.9 (3.5 to 8.2) <0.0001
 25% N2O 42.3 [40.7 to 45.9] 45.8 [42.8 to 52.6] 4.7 (3.3 to 6.1) <0.0001
 50% N2O 43.6 [40.9 to 47.9] 50.2 [43.8 to 52.8] 3.9 (2.6 to 5.2) <0.0001
P = 0.1244 (F = 3.659) P = 0.6321 (F = 0.9173) P = 0.8139 (F = 0.4118)
 Intubation No data No data No data N/A
 Incision 75.1 [66.9 to 84.5] 79.8 [72.2 to 92.7] 9.2 (5.2 to 13.2) <0.0001
 0% N2O 70.3 [62.8 to 79.8] 81.0 [75.5 to 96.8] 11.9 (9.1 to 14.7)U <0.0001
 25% N2O 80.3 [70.5 to 91.0] 86.8 [78.4 to 101.0] 8.4 (5.3 to 11.5)V <0.0001
 50% N2O 75.3 [70.8 to 92.3] 88.6 [75.0 to 98.8] 7.7 (5.1 to 10.2)W <0.0001
P = 0.0531 (F = 5.871) P = 0.6291 (F = 0.9268) P = 0.0307 (F = 6.968)V vs. U: 0.0668W vs. U: 0.0668
ANI, analgesia nociception index; BIS, bispectral index; CI, confidence interval; F, Friedman statistic; HR, heart rate; MAP, mean arterial blood pressure; NOL, nociception level index.

Fig. 4:
Alteration of each study parameter before and after nociceptive stimulation represented as average for 1-min prestimulus for baseline and average of the five values around the poststimulus peak (10 s before peak, peak and 10 s after peak)

Figure 2 shows the time course of the five parameters: the NOL index, the ANI index, the HR, the MAP and the BIS. Baseline values (the average of the last minute before stimulus) were compared to assess any intrinsic effect of N2O variation (Table 3). Friedman tests showed no significant differences between baseline values at 0, 25 and 50% for the NOL index (P = 0.3012), BIS (P = 0.1244) and MAP (P = 0.0531). However, baselines values were significantly different for the ANI index (P = 0.0063, 0 vs. 50%, adjusted P = 0.0647, 0 vs. 25%, adjusted P = 0.0076, 25 vs. 50%, adjusted P = 0.9999) and HR (P = 0.0003, 0 vs. 50%, adjusted P = 0.0027, 0 vs. 25%, adjusted P = 0.0007, 25 vs. 50%, adjusted P = 0.9999) (Table 3, prestimulation column).

Our study's primary objective was to show a 50% decrease in ΔANI between 0 and 50% N2O. This objective was not reached as ΔANI was 20.3 (25.9 to 14.7) at 0% N2O and 17.0 (24 to 10.1) at 50% N2O (NS), only a 16% decrease of ANI reaction to stimulation at 50% N2O.

Figure 2 also shows the poststimulus area under the curve (AUC, trapezoidal method) for each variable and at each N2O concentration (0, 25, 50%). Significant differences in poststimulation AUC exist between 0 and 25% and 0 and 50% N2O for the NOL (P = 0.0065, 0 vs. 50%, adjusted P = 0.0425, 0 vs. 25%, adjusted P = 0.0090, 25 vs. 50%, adjusted P = 0.9999) and the ANI indices (P < 0.0001, 0 vs. 50%, adjusted P = 0.0004, 0 vs. 25%, adjusted P < 0.0001, 25 vs. 50%, adjusted P = 0.6367), and 0 vs. 25% for HR (P = 0.0431, 0 vs. 50%, adjusted P = 0.3287, 0 vs. 25%, adjusted P = 0.0415, 25 vs. 50%, adjusted P = 0.9999), but not for MAP (P = 0.0531) and BIS (P = 0.1244).

Electrical tetanic stimulation elicited significant increases of the measured variables when they were evaluated as poststimulation peaks (see statistical analysis paragraph for explanation on the calculation of the peaks) (Fig. 4b, Table 3). Dotted lines in Fig. 4 indicate what would represent a clinically significant variation of the variable. For the NOL index, a value at least 25 was considered clinically meaningful to detect nociception (based on previous studies). For the ANI, a decrease below the value of 50 is generally seen as a clinical threshold for detecting nociception. For all other parameters (HR, BP, BIS), a classically reported increase of more than 10% of the prestimulation value was considered clinically significant.

Variations of each variable were analysed as absolute differences (Δ) and compared between N2O concentrations (0, 25 and 50%) (Fig. 4c, Table 3). Friedman tests were used to evaluate if variations were different between the three different N2O concentrations (0, 25 and 50%). NOL and HR showed significant differences between N2O concentration (P = 0.0035 and 0.0006, respectively). In both cases, 25 and 50% were statistically different from 0%, but 25 and 50% were not different (refer to Table 3 for details).

We also evaluated the effect of N2O on the NOL response to electrical tetanic stimulation in terms of percentage of patients with significant modification [(ΔNOLexp − ΔNOLCTRL)/ΔNOLCTRL × 100]. N2O induced a significant modification of the ΔNOL in 71% of the patients when treated with 50% N2O/O2 and 81% when treated with 25% N2O/O2 compared with 0%. The ratio of patients who showed variation following N2O inhalation was not significantly different between 25 and 50% N2O/O2 (Fisher test, P = 0.5541). The average posttetanic stimulus NOL index reduction was −35.3% [95% confidence interval (CI) −54.5 to −16.2%] with 50% N2O/O2, and −37.2% (95% CI −53.8 to −20.7%) with 25% N2O/O2. Both were significantly different from 0% N2O, using a Wilcoxon signed rank test (50% N2O, P = 0.0019; 25% N2O, P < 0.0001). The average posttetanic stimulus NOL index reduction between 50% N2O/O2 and 25% N2O/O2 was not significantly different (Mann–Whitney test, P = 0.6631).

For the ANI index, an antinociceptive effect through the ΔANI was observed in 55% when treated with 50% N2O/O2, and in 69%, when treated with 25% N2O/O2. The difference was NS between the two concentrations 25 and 50% N2O (Fisher test, P = 0.4173) compared with 0%. The average ANI score variation was 5.3% (95% CI −42% to +52.6%) with 50% N2O/O2, and −55.6% (95% CI −87.5% to −23.7%) with 25% N2O/O2. Only the variation seen with 25% N2O/O2 was statistically different from zero (50% N2O/O2, P = 0.8040; 25% N2O/O2, P = 0.0006), and variations of ANI for 25 vs. 50% were significantly different (Mann–Whitney test, P = 0.0145).


The current study represents an effort to quantify the antinociceptive effect of N2O during general anaesthesia using a combination of clinical parameters (HR and BP), and two newly developed intra-operative nociception monitors (ANI and NOL indices).

The results illustrated significant antinociceptive properties of 25 and 50% inhaled N2O revealed by ANI, NOL and HR, a reduction of 16% in ANI, a reduction of 31% in NOL reaction and a reduction of 51% in HR reaction at 50% N2O in response to tetanic electrical noxious stimulations but no significant difference between the antinociceptive effect produced by 50 or 25% inhaled N2O in O2.

The current study reported significant variations for each study parameter in response to a standardised24 electrical tetanic stimulation following 0, 25 and 50% N2O in O2 when looking at the poststimulation averaged peaks (Fig. 4, Table 3). This shows that all five variables assessed are able to detect intra-operative nociception. Significantly, there was no difference between the analgesic effect of 50 vs. 25% N2O (Figs. 2 and 4c). This agrees with previous reported studies in rodents where a low concentration of N2O was sufficient to induce antinociception.9,12

Our current results suggest that some clinically significant antinociceptive properties of N2O starting at low inhaled concentrations (25%) in anaesthetised patients exist. The concentration needed to obtain the antinociceptive effect is considerably lower than was previously thought, which also brings an important reduction in the pollution induced by N2O. Indeed, N2O is considered a greenhouse gas and a gas with a large environmental impact,25 leading some authors to strongly recommend the avoidance of N2O.26,27

The analgesic properties of N2O were described for the first time around 1800, but started to be used as a mixture with O2 (80% N2O/20% O2) in 1881.28 It has mild sedative and analgesic properties.28 The precise cellular mechanism carrying the anaesthetic/analgesic property of N2O remains elusive, although recent studies have suggested a noncompetitive NMDA receptor antagonist effect, and also potentiation of the gamma-aminobutyric acid receptor.6,29 It is very likely that N2O targets multiple postsynaptic sites through ligand-gated ion channels.5,6,29 Other than N2O, opioids are commonly used to improve anaesthesia and to produce antinociception during surgery. A disadvantage of opioids is that they seem, when given at high intra-operative doses, to be responsible for postoperative hyperalgesia.11,13,30–33 In rats, it was shown that intra-operative administration of 50% N2O in O2, or even lower concentrations, significantly reduced postoperative hyperalgesia associated with intra-operative opioid treatment.12 In humans, N2O was shown to decrease postoperative pain compared with remifentanil.34 However, it is not clear if N2O could have a similar effect as ketamine on postoperative sensitisation.11,35 Until the current study, it was not clear whether N2O had clinically significant antinociceptive properties during anaesthesia. Here, an experimental study design in anaesthetised patients with a nonoperator dependent stimulus (standardised tetanic electrical stimulation) demonstrated that N2O is able to significantly reduce the nociceptive response to this stimulation.

Also, the antinociceptive properties of N2O revealed in our current study carry weight to the argument for anaesthesiologists who are in favour of continuing N2O administration in the operating room setting, as low concentrations mean also less pollution.

In this study, several variables which assess the reduction of nociceptive responses to standardised stimuli were compared. This is because some techniques have recently been shown to have higher sensitivity and specificity to detect nociception under general anaesthesia than clinical signs such as HR and BP.18–21 In addition to the ANI and the NOL indices, we measured HR, BIS and MAP to quantify intra-operative nociception after stimulations. Indeed, under general anaesthesia, and following electrical tetanic stimulation, each parameter showed significant variations (Fig. 4b). The cardiovascular and haemodynamic response to a noxious stimulus is known to be associated with the antinociceptive property of the anaesthetics.36 This strongly supports the hypothesis that even under general anaesthesia, nociception occurs and needs to be controlled with appropriate analgesics.37

In this study, we compared NOL with ANI and found a number of subjects who had decreased intra-operative nociception measured by NOL compared with ANI were 71 and 81% vs. 50 and 69%, at 50 and 25% N2O. In addition, ΔNOL following the nociceptive stimulation seemed to be more clinically relevant than HR variations. The use of a nociceptive threshold for NOL (set at 25 from previous studies20,21) made it easier to use and notice changes after nociceptive stimulus in a clinical setting than using the delta HR in our study of 11.3 (95% CI 8.3 to 14.0) bpm for N2O at 0% with HRs varying from 66.8 (60.2 to 78.0) to 80 (69.8 to 92.6) bpm before and after the tetanic stimulation, respectively. This mild increase in HR, even if statistically significant, might not even be seen by the anaesthesiologist in the operating room as it is a very slight variation. Using a more sensitive and specific index, such as NOL with a set nociceptive threshold of 25, may prove to be of more clinical relevance in our daily anaesthesia practice.

The ANI baseline values increased between 0, 25 and 50% N2O. This was not problematic but a confusing result was the absence of significant difference in the ΔANI in response to the electrical stimulations at different N2O concentrations. We were not able to discern a decrease of the ΔANI in a situation of 50 or 25% N2O (Fig. 4c and Table 3). This makes ANI, under the same conditions used for HR and NOL, less robust than the combined use of HR and NOL to precisely characterise nociception after tetanic stimulation. Similarly, the ANI score was considerably reduced, even without any stimulation, with a median score of 63.2 (0% N2O), 77.0 (25% N2O) and 69.6 (50% N2O), and with, nine, three and two patients below the line of 50, indicating poor analgesia where no stimulation occurred, which in turn decreased the specificity of ANI for detecting nociception. Moreover, all other variables were stable and free from any indication of nociception (HR, BP, NOL, BIS). Conversely, if we compared previous studies, baseline ANI scores were fluctuating between 100 and 6838 and even lower.39 Also, the positive predictive value of the ANI seemed to be somewhat low.40 NOL and HR showed better results to characterise and identify the antinociceptive effect of N2O.

There are some limitations to our study. The inclusion of patients scheduled for cytoreduction surgery and HIPEC via laparotomy in this study may be seen as a limitation because intraperitoneal chemotherapy might induce physiological changes which interfere with measured variables. Nevertheless, all the procedures were completed during the cytoreduction phase and prior to the start of the intraperitoneal chemotherapy, consequently preventing any impact on the results. A total of three tetanic stimulations were applied to the forearm of the patient. It could be argued that one stimulation may have influenced the others located in the same area of the patient. However, we waited for at least 25 min between stimulations to significantly decrease the risk of carry-over effect of one stimulation to the next (Fig. 1). The variations of baselines values prior to the stimulation for ANI and HR reported higher ANI and lower HR at 25% that occurred after the stimulus at 50%. This difference in baseline values, even if small and not really clinically relevant, was statistically significant and might be seen as a carry-over effect of the 50% N2O that was applied before the 25%, thus increasing the analgesic effect of N2O when administered at 25%. However, baseline values were not different with the NOL index.

The variation of desflurane concentrations at the time the stimulus was applied might be seen as one limitation as recent reports reinforced the likelihood that desflurane might have some analgesic effect on its own. Here, when N2O was used, the desflurane concentration was lower (at 25 and 50% N2O concentrations) than without N2O (0%; Table 1). So, if desflurane would have had an impact on our primary outcome that was to measure the analgesic effect of N2O, this desflurane effect would have been stronger in the 0% N2O group, which would have minimised the difference in terms of NOL and ANI reactivity in the 0% N2O group vs. the 25 and 50% N2O groups. As this study did show a significant difference between 0% and 25 or 50% N2O in terms of antinociception effect of N2O, whereas the desflurane concentration was lower in the N2O 25 and 50%, this means that this difference would have been even higher if we had kept the desflurane concentration similar in the three groups. For safety reasons, we decided to decrease the desflurane concentration in the two N2O groups based on the BIS values and to avoid major and possibly unsafe changes in haemodynamics.

Epidural analgesia, with 3 ml of lidocaine 2% as a test dose and a total epidural dose of 8 ± 0.9 ml given prior to first incision, might have influenced some study variables such as NOL and ANI. All epidural catheters were placed below T9. It has been shown that epidural block does not influence the response of the NOL to a standardised tetanic electrical stimulation.20,21 The current results confirm that HR, BIS and NOL baseline values assessed in this study were not significantly changed when the epidural space was loaded with local anaesthetic, but ANI index did vary significantly (Table 2).

In conclusion, this study suggests that clinically significant antinociceptive properties of N2O are better identified with a combination of HR and the NOL index in anaesthetised patients undergoing laparotomy. Our results illustrated that low concentrations of N2O (25%) were as effective as higher concentrations (50%) in achieving a significant antinociceptive effect. These findings may help to decrease the negative effects of using higher concentrations of N2O, including its various side effects and its adverse environmental impact.

Acknowledgements relating to this article

Assistance with the study: special thanks to NG, RN and Research Coordinator in the Department of Anaesthesiology and Pain Medicine and who was in charge of organising this study at Maisonneuve-Rosemont Hospital. Thanks to Kyle Vaughn Roerick for his English language edition of this article.

Financial support and sponsorship: this study was sponsored by American Air Liquide Inc., Delaware Research and Technology Center (Newark, Delaware, USA) through an Independent Investigator Initiated Trial grant and a contract between the company and the Research Center of Maisonneuve-Rosemont Hospital/CEMTL and the principal investigator PR, Associate Professor at the University of Montreal. Medasense Biometrics Ltd graciously loaned the PMD200 device and offered the supplies for the NOL index evaluation. The Physiodoloris device for the ANI index evaluation is the property of the department of Anaesthesiology of Maisonneuve-Rosemont Hospital. This study was also supported by the Department of Anaesthesiology and Pain Medicine of the Maisonneuve-Rosemont Hospital, University of Montreal.

Conflicts of interest: PR is a member of the advisory board of the company Medasense Biometrics Ltd and, as such, he has received an honorarium since 2015. He has also received honoraria in the last 5 years as a consultant from Abbvie, Medtronic, Biosyent, Edwards, Avirpharma for lectures. No other conflicts of interest.

Presentation: none.


1. Jain K, Sethi SK, Damor M, et al. Effects of inhaled nitrous oxide on the induction dose and time requirements of propofol: a prospective, randomized, double-blind study. Anesth Essays Res 2017; 11:174–180.
2. Fujinaga M, Maze M. Neurobiology of nitrous oxide-induced antinociceptive effects. Mol Neurobiol 2002; 25:167–189.
3. Maze M, Fujinaga M. Recent advances in understanding the actions and toxicity of nitrous oxide. Anaesthesia 2000; 55:311–314.
4. Ohashi Y, Guo T, Orii R, et al. Brain stem opioidergic and GABAergic neurons mediate the antinociceptive effect of nitrous oxide in Fischer rats. Anesthesiology 2003; 99:947–954.
5. Jevtovic-Todorovic V, Todorovic SM, Mennerick S, et al. Nitrous oxide (laughing gas) is an NMDA antagonist, neuroprotectant and neurotoxin. Nat Med 1998; 4:460–463.
6. Mennerick S, Jevtovic-Todorovic V, Todorovic SM, et al. Effect of nitrous oxide on excitatory and inhibitory synaptic transmission in hippocampal cultures. J Neurosci 1998; 18:9716–9726.
7. Yamakura T, Harris RA. Effects of gaseous anesthetics nitrous oxide and xenon on ligand-gated ion channels. Comparison with isoflurane and ethanol. Anesthesiology 2000; 93:1095–1101.
8. Stiglitz DK, Amaratunge LN, Konstantatos AH, et al. Intra-operative nitrous oxide as a preventive analgesic. Anaesth Intensive Care 2010; 38:890–893.
9. Ben Boujema M, Laboureyras E, Pype J, et al. Nitrous oxide persistently alleviates pain hypersensitivity in neuropathic rats: a dose-dependent effect. Pain Res Manag 2015; 20:309–315.
10. Bessiere B, Laboureyras E, Chateauraynaud J, et al. A single nitrous oxide (N2O) exposure leads to persistent alleviation of neuropathic pain in rats. J Pain 2010; 11:13–23.
11. Echevarria G, Elgueta F, Fierro C, et al. Nitrous oxide (N2O) reduces postoperative opioid-induced hyperalgesia after remifentanil-propofol anaesthesia in humans. Br J Anaesth 2011; 107:959–965.
12. Richebe P, Rivat C, Creton C, et al. Nitrous oxide revisited: evidence for potent antihyperalgesic properties. Anesthesiology 2005; 103:845–854.
13. Wehrfritz A, Schaefer S, Troester A, et al. A randomized phase I trial evaluating the effects of inhaled 50–50% N2O–O2 on remifentanil-induced hyperalgesia and allodynia in human volunteers. Eur J Pain 2016; 20:1467–1477.
14. Chan MT, Wan AC, Gin T, et al. Chronic postsurgical pain after nitrous oxide anesthesia. Pain 2011; 152:2514–2520.
15. Richebe P, Capdevila X, Rivat C. Persistent postsurgical pain: pathophysiology and preventive pharmacologic considerations. Anesthesiology 2018; 129:590–607.
16. Richebe P, Julien M, Brulotte V. Potential strategies for preventing chronic postoperative pain: a practical approach: continuing professional development. Can J Anaesth 2015; 62:1329–1341.
17. Loeser JD, Treede RD. The Kyoto protocol of IASP basic pain terminology. Pain 2008; 137:473–477.
18. Edry R, Recea V, Dikust Y, et al. Preliminary intra-operative validation of the nociception level index: a noninvasive nociception monitor. Anesthesiology 2016; 125:193–203.
19. Martini CH, Boon M, Broens SJ, et al. Ability of the nociception level, a multiparameter composite of autonomic signals, to detect noxious stimuli during propofol-remifentanil anesthesia. Anesthesiology 2015; 123:524–534.
20. Renaud-Roy E, Stockle PA, Maximos S, et al. Correlation between incremental remifentanil doses and the Nociception Level (NOL) index response after intra-operative noxious stimuli. Can J Anaesth 2019; 66:1049–1061.
21. Stockle PA, Julien M, Issa R, et al. Validation of the PMD100 and its NOL Index to detect nociception at different infusion regimen of remifentanil in patients under general anesthesia. Minerva Anestesiol 2018; 84:1160–1168.
22. Issa R, Julien M, Decary E, et al. Evaluation of the analgesia nociception index (ANI) in healthy awake volunteers. Can J Anaesth 2017; 64:828–835.
23. Logier R, Jeanne M, De Jonckheere J, et al. PhysioDoloris: a monitoring device for analgesia/nociception balance evaluation using heart rate variability analysis. Conf Proc IEEE Eng Med Biol Soc 2010; 2010:1194–1197.
24. Rantanen M, Ypparila-Wolters H, van Gils M, et al. Tetanic stimulus of ulnar nerve as a predictor of heart rate response to skin incision in propofol remifentanil anaesthesia. Br J Anaesth 2007; 99:509–513.
25. Ravishankara AR, Daniel JS, Portmann RW. Nitrous oxide (N2O): the dominant ozone-depleting substance emitted in the 21st century. Science 2009; 326:123–125.
26. Ryan S, Sherman J. Sustainable anesthesia. Anesth Analg 2012; 114:921–923.
27. Ryan SM, Nielsen CJ. Global warming potential of inhaled anesthetics: application to clinical use. Anesth Analg 2010; 111:92–98.
28. Schonherr ME, Hollmann MW, Graf B. Nitrous oxide. Sense or nonsense for today's anaesthesia. Anaesthesist 2004; 53:796–812.
29. Dzoljic M, Van Duijn B. Nitrous oxide-induced enhancement of gamma-aminobutyric acid A-mediated chloride currents in acutely dissociated hippocampal neurons. Anesthesiology 1998; 88:473–480.
30. Angst MS, Koppert W, Pahl I, et al. Short-term infusion of the mu-opioid agonist remifentanil in humans causes hyperalgesia during withdrawal. Pain 2003; 106:49–57.
31. Bessiere B, Richebe P, Laboureyras E, et al. Nitrous oxide (N2O) prevents latent pain sensitization and long-term anxiety-like behavior in pain and opioid-experienced rats. Neuropharmacology 2007; 53:733–740.
32. Celerier E, Rivat C, Jun Y, et al. Long-lasting hyperalgesia induced by fentanyl in rats: preventive effect of ketamine. Anesthesiology 2000; 92:465–472.
33. Rivat C, Laulin JP, Corcuff JB, et al. Fentanyl enhancement of carrageenan-induced long-lasting hyperalgesia in rats: prevention by the N-methyl-d-aspartate receptor antagonist ketamine. Anesthesiology 2002; 96:381–391.
34. Kim MK, Yi MS, Kang H, et al. Effects of remifentanil versus nitrous oxide on postoperative nausea, vomiting, and pain in patients receiving thyroidectomy: propensity score matching analysis. Medicine (Baltimore) 2016; 95:e5135.
35. Nistal-Nuno B, Freire-Vila E, Castro-Seoane F, et al. Preoperative low-dose ketamine has no preemptive analgesic effect in opioid-naive patients undergoing colon surgery when nitrous oxide is used – a randomized study. F1000Res 2014; 3:226.
36. Antognini JF, Carstens E. In vivo characterization of clinical anaesthesia and its components. Br J Anaesth 2002; 89:156–166.
37. Lichtner G, Auksztulewicz R, Velten H, et al. Nociceptive activation in spinal cord and brain persists during deep general anaesthesia. Br J Anaesth 2018; 121:291–302.
38. Gruenewald M, Herz J, Schoenherr T, et al. Measurement of the nociceptive balance by Analgesia Nociception Index and Surgical Pleth Index during sevoflurane-remifentanil anesthesia. Minerva Anestesiol 2015; 81:480–489.
39. Ledowski T, Averhoff L, Tiong WS, et al. Analgesia Nociception Index (ANI) to predict intra-operative haemodynamic changes: results of a pilot investigation. Acta Anaesthesiol Scand 2014; 58:74–79.
40. Boselli E, Daniela-Ionescu M, Begou G, et al. Prospective observational study of the noninvasive assessment of immediate postoperative pain using the analgesia/nociception index (ANI). Br J Anaesth 2013; 111:453–459.
Copyright © 2021 European Society of Anaesthesiology and Intensive Care. Unauthorized reproduction of this article is prohibited.