Abstract
Conscious patients undergoing regional anaesthesia are able to interact with caregivers and indicate when they are feeling pain and/or anxiety during surgery. Unconscious patients on the other hand, not only in the operating room but also in intensive care, need antinociception when a nociceptive stimulus is applied. In these patients, the dose of antinociceptive drugs cannot be adjusted or the effect of locoregional anaesthesia supplementing general anaesthesia cannot be determined by simply asking the patient. Thus, the clinician needs to rely on monitors of the nociception/antinociception (N/AN) balance to be able to provide adequate, that is personalized, antinociception.
Introduction
Conscious patients undergoing regional anaesthesia are able to interact with caregivers and indicate when they are feeling pain and/or anxiety during surgery. Unconscious patients on the other hand, not only in the operating room but also in intensive care, need antinociception when a nociceptive stimulus is applied. In these patients, the dose of antinociceptive drugs cannot be adjusted or the effect of locoregional anaesthesia supplementing general anaesthesia cannot be determined by simply asking the patient. Thus, the clinician needs to rely on monitors of the nociception/antinociception (N/AN) balance to be able to provide adequate, that is personalized, antinociception.
Defining Analgesia and Nociception in Non-conscious States
The reader will have noticed that the terms ‘nociception’ and ‘antinociception’ rather than ‘pain’ and ‘analgesia’ are being used. The term ‘analgesia’ refers to pain, which has been defined by WHO as ‘an unpleasant sensory and emotional experience associated with actual or potential tissue damage, or described in terms of such damage’. In other words, it describes how a conscious being integrates a nociceptive stimulus into a negative emotional experience. Therefore, the term ‘analgesia’ cannot be used in non-conscious states such as those induced by hypnotic drugs during general anaesthesia, and the term ‘antinociception’ has to be used instead. But given the state of unconsciousness induced by anaesthetic drugs, how can we quantify nociception after a noxious stimulus has been applied to the mechanical or biochemical sensors in the human body that generate nociceptive signals? And how can we account for inter-individual variability?
Antinociception has traditionally been quantified with non-specific measurements of the clinical response to noxious stimuli such as loss of verbal response, immobility and absence of hypertension and tachycardia, which are indicators of loss of consciousness, loss of motor response or loss of haemodynamic reactivity, respectively. How can antinociception (i.e. absence of verbal response, immobility, and absence of hypertension and tachycardia after a noxious stimulus) be provided?
One way to prevent so-called ‘purposeful movement’ is to administer a potent inhaled anaesthetic. At 1MAC or 1.3MAC, they ensure 50% or 95% of the patients will not move (see Chapter 6 for more details). The end-expired partial pressure needed to attain the same probability of response suppression is decreased by GABA-acting agents (benzodiazepines, propofol), NMDA antagonists (ketamine), α2-adrenergic agonists (clonidine or dexmedetomidine) and of course the potent opioids (the drugs most commonly used to provide antinociception during general anaesthesia). All these drugs have in common that they attenuate the autonomic nervous system (ANS) response to nociception that we usually observe as tachycardia and hypertension. Regional analgesia prevents noxious stimuli from reaching the central nervous system, thus also attenuating the ANS response. This chapter will mainly focus on monitoring the effects of the nociception/antinociception balance via the ANS response.
Quantitative and Modelling Information
Administration of Opioids: The Pharmacokinetic/Pharmacodynamic Models
Opioids can be administered as intermittent boluses, by manually adjusting a continuous infusion or by using a target controlled infusion (TCI) device, with the target being either the plasma (Cp) or effect-site concentration (Ce). They are often titrated towards haemodynamic goals such as heart rate and blood pressure. TCI systems are based upon pharmacokinetic/pharmacodynamic (PK/PD) models. Targeting effect-site rather than plasma concentration takes the hysteresis into account. The effect-site concentration is by definition directly correlated with the effect. TCI is intended to improve bias (‘offset’) and accuracy (‘togetherness’), and thus helps ensure more stable and predictable opioid concentrations, but biological variability will continue to place a limit on the accuracy and bias of these devices. The benefits related to the use of these devices are still a matter of debate since, from a PD point of view, the optimal target is not the concentration predicted by PK/PD models but an individual’s response that should be measured, or at least estimated, in each patient. Based on the measurement of the effect, the concentration of drugs can be adjusted until it attains the targeted effect.
Administering an opioid dose to an individual patient that is likely to provide a certain concentration and a certain effect based on a population model goes a long way towards achieving personalized, ‘optimized’ or ‘tailored’ anaesthesia, but ultimately and optimally we must also measure the clinical effect to ascertain that it has indeed been obtained and use the clinical effect in the individual to help steer further drug administration.
How to Achieve Personalized Analgesia
Personalized anaesthesia may be achieved by using monitors that measure any of the several components of anaesthetic depth such as immobility provided by neuromuscular blockers [3, 4] (with a neuromuscular transmission monitor), depth of hypnosis (with monitors like the bispectral index (BIS), state entropy (SE), or any other EEG derived index [5, 6]), and the N/AN balance (with several commonly called ‘analgesia monitors’). ‘Analgesia monitors’ actually is a misnomer, ‘antinociception monitors’ or ‘N/AN balance monitors’ would be the more correct terms. Most analgesia monitors are based on quantifying measurable responses directly related to the variations in tonus of the ANS (sympathetic and parasympathetic activity) in relation to changes occurring in the N/AN balance [7, 8]. These monitors use various physiological signals: heart rate, heart rate variability, blood pressure, skin conductance, plethysmograph waveform, pupil diameter or sometimes a combination of them (composite indicators). They can also integrate signals coming from EEG activity. These monitors are reviewed in the remainder of this chapter.
How to Measure Antinociception
Monitors that Use Signals from the Cerebral Cortex as an Indication of the Nociception/Antinociception Balance
Several monitors derive an ‘analgesia index’ from cortical activation after a nociceptive stimulus, assuming this response to reflect ‘a state of antinociception’. The indices are dimensionless numbers, and are the results from complex signal analysis and indicator extraction (see Chapter 5 for details). The qNOX index for example results from the analysis of four EEG spectral bands between 0.5 and 44 Hz. Clinical studies have shown that the qNOX was mildly correlated with the effect-site concentration of remifentanil and with the probability of response to nociception [9].
Monitoring Sympathetic and Parasympathetic Activity as a Surrogate for the Nociception/Antinociception Balance
The sympathetic and parasympathetic systems are affected by the N/AN balance and simultaneously affect several organ systems, the responses of which can be used to help quantify this balance.
Analgesia/Nociception Index®
The Analgesia/Nociception Index® (ANI) (MDoloris Medical Systems, Lille, France) is a wavelet transform-based heart rate variability index with values between 0 and 100, and is a measure of the relative parasympathetic tone [8, 10]. The ANI is based on the measurement of the impact of ventilation on the R-R interval of the electrocardiogram, and allows a qualitative and quantitative measurement of heart rate variability (HRV), in particular in the high-frequency power (0.15–0.5 Hz) (see Fig. 8.1) [11]. The ANI was primarily developed to assess the N/AN balance during general anaesthesia [11]. In clinical practice, high ANI values (> 55–60) indicate parasympathetic predominance, i.e. adequate antinociception, whereas low ANI values (< 50) indicate sympathetic predominance, i.e. nociception [8, 10]. Briefly, a low ANI denotes insufficient ‘analgesia’.
Fig. 8.1 Mean centered, normalized and band pass-filtered R-R series in two different levels of nociception/antinociception balance (upper panel: surgical stimulus in the case of adequate nociception/antinociception balance, lower panel: surgical stimulus in the case of inadequate antinociception (lower panel). Figure reproduced from [8] (with permission).
The performance of ANI has been evaluated in several studies both in children and in adults during general anaesthesia.
In children receiving desflurane titrated to a BIS of 50, a standardized nociceptive stimulus (ulnar tetanic stimulation) was applied after stepwise reductions of a remifentanil infusion (from 0.2 to 0.04 µg/kg/min). The ANI was always lower after than before stimulation, and lower with lower doses of remifentanil [12]. In another study in children, the ANI was used to predict failure of regional anaesthesia (defined by an increase in heart rate by > 10% within 2 min after skin incision). Area under the receiver-operating curve (ROCAUC) to identify regional anaesthesia failure was 0.747, with a 79% sensitivity and 62% specificity for ANI values <51 [13].
While maintaining the BIS within 30–60 with propofol in adults, a stepwise increase of the remifentanil Ce from 0 to 2 and 4 ng/mL followed by tetanic stimulation of the ulnar nerve resulted in a median ANI of 24, 30 and 13, respectively. Prediction of movement by ANI following stimulation was poor, with a prediction probability (Pk) 0.41; Pk is a statistical measure analogous to ROC AUC, with 0.5 indicating a prediction probability equal to flipping a coin [14]. Similar results were observed in patients under a sevoflurane and fentanyl general anaesthesia: mean ANI decreased from 52 to 33 (p<0.005) after airway manipulation and from 63 to 38 (p<0.001) after skin incision (indicating that the ANI reflected the effect of the noxious stimulus well), but the prediction probability for a 10% increase of heart rate and or systolic blood pressure was low (Pk of 0.61 and 0.59, respectively) [15]. During total knee replacement, the ANI ‘early detected’ haemodynamic reactivity (defined as a 20% increase in heart rate or systolic blood pressure following intraoperative stimulation) (ROC AUC = 0.92), with 80% sensitivity and 88% specificity at the ANI threshold of 63 [16].
However, if the objective is to optimize analgesic drug titration, it is important to be able to predict whether a decrease in ANI (= poor N/AN balance) will occur after a noxious stimulus rather than merely being able to detect it post factum. This has been reported on in three studies, with various results. During suspension laryngoscopy using propofol/remifentanil or desflurane/remifentanil general anaesthesia, ANI was a good predictor of haemodynamic reactivity (ROC AUC 0.88 and 0.77 with propofol/remifentanil and desflurane/remifentanil, respectively) [17, 18]. The predictive performance of dynamic variations of ANI within 1 min (∆ANI) was even better (ROC AUC = 0.90, with 85% sensitivity and specificity for a variation of –19% in 1 min) [18]. However, when fentanyl rather than remifentanil was used during surgery, the prediction probability of the ANI to predict more than 10% increase in heart rate or systolic blood pressure was poor (PK 0.61 and 0.59, respectively)[15]. The PK/PD differences in remifentanil (a short-acting opioid with short elimination half-life of < 10 min) and fentanyl (a long-acting opioid with long elimination half-life of > 10 min) may explain these discrepancies [19, 20].
ANI values at the end of surgery and immediately before extubation may help predict immediate postoperative pain [21]. A negative linear relationship was observed between ANI immediately before extubation and NRS [0–10 numerical rating scale (NRS)] on arrival in the PACU. In an observational study performed in 200 patients undergoing ear, nose and throat or lower limb orthopaedic surgery with general anaesthesia using an inhalational agent and remifentanil, the ANI prior to extubation predicted the occurrence of pain (NRS>3) after arrival in PACU well (ROC AUC = 0.89, with 86% sensitivity and specificity for ANI < 50) [21]. In 120 patients undergoing elective laparoscopic cholecystectomy, Szental et al reported that the morphine administration guided by the ANI instead of by clinical signs did not decrease the rate of moderate/severe pain or the use of rescue analgesia in PACU [22]. However, morphine might not have been the optimal choice for ANI guidance considering its slow onset and long acting PK/PD properties. When fentanyl boluses were titrated to maintain the ANI ≥ 50 instead of using standard clinical administration in 50 patients undergoing lumbar discectomy or laminectomy, postoperative NRS pain scores were lower (1.3 on average), rescue fentanyl administration was less frequent and nausea scores lower [23].
To summarize, the ANI is a continuous 0–100 index of the relative (i.e. parasympathetic /sympathetic) ANS tone. ANI values decrease after insufficiently blunted nociception and its dynamic variations that may help predict haemodynamic reactivity. It intends to optimize intraoperative titration of opioids, but it requires further clinical validation. ANI monitoring cannot be used in patients with cardiac arrhythmia (because the R-R interval is erratic), patients suffering from ANS dysfunction (because this might alter ANI values), and may be impaired in patients receiving inotropic drugs or β-blockers, although this has to be further evaluated.
Cardiovascular Depth of Analgesia®
The algorithm of the CARdiovascular DEpth of ANalgesia® or CARDEAN index (Alpha-2 Ltd, Lyon, France) predicts hypertension followed by tachycardia based on beat-to-beat analysis of blood pressure and heart rate. The index (0–100%) measures the degree of inhibition of the baroreflex caused by an inadequate N/AN balance [8]. Higher values indicate high levels of baroreflex inhibition, defined as a paradoxical decrease in the R-R interval following an increase in systolic blood pressure and correspond to nociception, whereas an increase in the R-R interval is expected after an increase in systolic blood pressure in the presence of adequate analgesia (see Fig. 8.2). In a retrospective study in 40 ASA 1–2 patients undergoing knee surgery under general anaesthesia without muscle relaxant, the CARDEAN predicted intraoperative movement with high accuracy (ROC AUC = 0.98, with 100% sensitivity and 95% specificity for a CARDEAN index > 60). When propofol was used to maintain the BIS at 40–60 in patients undergoing colonoscopy, titrating alfentanil to the CARDEAN index rather than to conventional signs (e.g. tachycardia, hypertension or movement) decreased the incidence of unexpected movements by 51% [34]. When propofol (titrated to a BIS of 40–60) and remifentanil were used to provide general anaesthesia for spinal disc repair [24], skin incision increased CARDEAN values if Ce remifentanil was 2 ng/mL, but not when it was 4 ng/mL.
Fig. 8.2 R-R and systolic blood pressure (SBP) series at two different levels of the nociception/antinociception balance; A) surgical stimulus in case of adequate anti-nociception (upper panel), B) surgical stimulus in the case of inadequate antinoci-ception (lower panel).
Summarized, preliminary data indicate that the CARDEAN index can be related to the N/AN balance; further clinical validation is warranted.
Pupillometry
The pupil diameter results from the balance between the sympathetic dilator tone and the parasympathetic constrictor tone [25]. It can change in many different situations, both in healthy subjects and in several disease states, including brain trauma. During anaesthesia, the pupil diameter changes both after the administration of certain drugs, especially – but not exclusively – after opioid administration and, in a reflex manner, after noxious stimulation. A portable infra-red pupillometry device can be used to measure pupil size and pupillary reflexes, in particular the pupillary dilation reflex (PDR) in response to a homogeneous noxious tetanic stimulation during anaesthesia [8, 25, 26].
In anaesthetized volunteers receiving a constant tetanic electrical stimulus, increasing alfentanil concentrations impaired the PDR but had no effect on the pupillary light reflex (PLR) [27]. The relationship between alfentanil concentrations and the post-stimulus change in pupil size was exponential. The authors concluded that the dilation of the pupil in response to a noxious stimulus was a measure of the A/AN balance in anaesthetized patients.
Several studies attest to the performance of this technique to measure the N/AN balance. During TCI with propofol and increasing concentrations of remifentanil (0 to 5 ng/mL), pupil size after a constant tetanic stimulus decreased from 1.55±0.72 mm to 0.01 ± 0.03 mm with increasing Ce remifentanil; pupil dilation and remifentanil concentration were inversely correlated [28]. The decrease in pupil response to a nociceptive stimulus correlated better with Ce remifentanil than with haemodynamics or BIS measurements.
In children receiving anaesthesia with sevoflurane and alfentanil, pupillary dilation after skin incision was a more sensitive measure of noxious stimulation than haemodynamic variables and BIS values [29]. The pupillary diameter increased by 200 ± 40 % within 1 min after skin incision, with a rapid return to pre-incision values after alfentanil administration.
In 80 female patients undergoing vacuum aspiration while receiving propofol and remifentanil TCI, the PDR amplitude (mean ± SD) was significantly (p<0.001) greater in movers (2.0 ± 1.2 mm) than in non-movers (0.6 ± 0.7 mm) following a homogeneous noxious stimulation (standardized tetanic stimulation) [30]. The performance of PDR amplitude in predicting movement was good, with an AUC ROC of 0.90 (95% CI 0.83–0.96). The PDR amplitude associated with a 50% and 95% probability of non-movement after noxious stimulation was 1.39 mm (95% CI, 0.96–2.20) and 0.29 mm (95% CI, 0.17–0.55). The authors concluded PDR amplitude monitoring could help optimize opioid administration during general anaesthesia.
Pupillometry has been used to assess the effectiveness of regional anaesthesia combined with general anaesthesia in both adults [31] and children [13]. In 24 adult patients undergoing elective foot or ankle surgery under general anaesthesia combined with a popliteal sciatic nerve block, the median [interquartile] PDR response to standardized lower limb tetanic stimulation was blunted to only 2% [1–4] in the blocked leg, whereas a 17% [13–24] increase was observed in the non-blocked leg (p<0.01) [31]. In 58 children undergoing elective surgery with combined general anaesthesia (with sevoflurane) and regional anaesthesia (central or peripheral nerve blocks), the effectiveness of regional anaesthesia was assessed using the pupillary diameter. A PDR cut-off value of > 4.2 mm identified regional anaesthesia failure (defined by a rise in heart rate ≥ 10% from baseline within 2 min after skin incision) with 58% sensitivity and 79% specificity.
Finally, the PDR has been used to assess immediate postoperative analgesia. In 100 patients, postoperative pain intensity was assessed shortly after arrival in the post-anaesthesia care unit using a 0–5 verbal rating scale (VRS) where zero equalled no pain, and five extreme pain. VRS values were linearly correlated with PDR (ρ = 0.88, p<0.0001). The threshold value of PDR corresponding to the highest accuracy to have VRS>1 was 23%, with 91% sensitivity and 98% specificity.
In summary, pupillometry seems to be an accurate marker of the N/AN balance, with the pupil diameter increasing with nociception and the PDR decreasing in cases of adequate antinociception. Pupillometry may help assess the efficacy of regional anaesthesia if combined with general anaesthesia, both in adults and in children. In the immediate postoperative period, a PDR value > 23% after noxious stimulation in the awake patient may accurately predict whether (s)he will be suffering from pain. The optimal thresholds to guide intraoperative analgesia need to be determined. The technique has some drawbacks. Pupillometry is not a continuous measurement – intermittent tetanic stimulations are required to measure PDR. In addition, its use may be limited when the eye of the patient is not accessible due to the nature of the surgery (e.g. head and neck or neurosurgery, steep Trendelenburg position) (see Fig. 8.3).
Skin Conductance
Skin conductance (SC) measures the modulation of electrical conductivity of the skin due to sweat production related to sympathetic activity [7, 8]. Fluctuations in SC in response to noxious stimulation depend on the level of antinociception present. One advantage of SC as compared with HRV or other measures derived from sympathetic nervous system activity is that the neurotransmission to the sweat glands is not adrenergic and thus is not influenced by the administration of β-blockers. The monitor measures SC changes from peak and trough values obtained over time (see Fig. 8.4) via three electrodes placed on palmar or plantar skin, and calculates the number of fluctuations in skin conductance (NFSC) and the corresponding AUC (expressed in µS) [32]. After sympathetic activation, e.g. after intraoperative noxious stimulation, the NFSC increases, which may indicate inadequate analgesia [32].
