BJA Advance Access originally published online on January 23, 2006
British Journal of Anaesthesia 2006 96(3):367-376; doi:10.1093/bja/ael005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
PAIN |
Novel multiparameter approach for measurement of nociception at skin incision during general anaesthesia
,
1 Department of Anaesthesia, Tampere University Hospital, PO Box 2000, FIN-33521 Tampere, Finland. 2 Medical School, University of Tampere, FIN-33014, Finland. 3 VTT Information Technology, PO Box 1206, FIN-33101 Tampere, Finland. 4 GE Healthcare Finland, PO Box 900, FIN-00031 GE, Finland. 5 Department of Anaesthesia and Intensive Care, Helsinki University Central Hospital, PO Box 340, FIN-00029 HUS, Finland
* Corresponding author. E-mail: arvi.yli-hankala{at}uta.fi
Accepted for publication December 12, 2005.
 |
Abstract |
|---|
 TopFootnotes Abstract IntroductionPatients and methodsResultsDiscussionAppendixReferences |
|---|
Background. Direct indicators for the evaluation of the nociceptive–anti-nociceptive balance during general anaesthesia do not exist. The aim of this study was to combine physiological parameters to obtain such an indicator.
Methods. Fifty-five females scheduled for surgery under general anaesthesia combining target-controlled infusions of propofol and remifentanil were studied. Propofol was given to maintain state entropy (SE) at 50 and remifentanil was targeted at 1, 3 or 5 ng ml–1. The patients' reactions and clinical signs of nociception, remifentanil levels and estimation of noxious intensity of incision were combined into a clinical score [Clinical Signs–Stimulus–Antinociception (CSSA)] to evaluate the nociceptive–anti-nociceptive balance. ECG, photoplethysmography (PPG), response entropy (RE) and SE were recorded from 60 s before to 120 s after skin incision. Differences between post- and pre-incision values of heart rate variability (HRV), PPG and pulse transition time related parameters were analysed off-line to evidence the best predictors of CSSA. Those best predictors of CSSA served to develop a response index of nociception (RN), scaled from 0 to 100. This index was further tested in 10 additional patients.
Results. HRV, RE, RE–SE and PPG variability were the best predictors of CSSA. The prediction probability of RN at predicting CSSA was 0.78. RN response was higher after larger incision, in movers and with lower remifentanil concentrations.
Conclusions. The empirically developed algorithm of RN leads to an index that seems to adequately estimate the nociceptive–anti-nociceptive balance at skin incision during general anaesthesia. In the future, CSSA may serve as a reference for studies investigating methods aimed at evaluating this pharmacodynamic component of anaesthesia.
Keywords: anaesthetics i.v., propofol; analgesics opioid, remifentanil; blood, flow, peripheral; brain, electroencephalography; heart, heart rate; monitoring, respiratory sinus arrhythmia
|
Introduction |
|---|
|
TopFootnotesAbstractIntroductionPatients and methodsResultsDiscussionAppendixReferences |
|---|
Pain is a descriptor of a conscious, emotional, private experience, generated by a wide variety of events. Tissue damage-associated acute pain consists of both physiological and psychological characters. During general anaesthesia, the conscious experience of pain disappears. Hence, in an anaesthetized patient it is not appropriate to call ‘pain’ the physiological reaction to tissue aggression by surgery. Instead, surgical stimulus with its immediate consequences on system function—without the higher order processing that is defined as consciousness—is recognized with the use of the word ‘nociception’.1 This term covers all activity caused by tissue damage from the peripheral sensory afferent to the brainstem level, where nociception may be manifested as cardiovascular or hormonal responses.1 2 In the present text, the term nociception will stand for the physiological reactions of the nervous system to surgical stimulation during unconsciousness induced by general anaesthesia and the term nociceptive–anti-nociceptive balance denotes the situation during which simultaneous opposite effects of nociceptive stimulation and anti-nociceptive (analgesic) medication on the physiology of an anaesthetized patient occur.
During surgery, the intensity of the autonomic or motor response to a nociceptive stimulus is dependent on the nociceptive–anti-nociceptive balance. Nociception can be attenuated with sufficient amount of anti-nociceptive medications. Hypnotic and anti-nociceptive components of balanced anaesthesia are partly interconnected: high concentrations of hypnotics lessen the effects of nociceptive stimulation,3 and a strong anti-nociceptive component of anaesthesia decreases the need for hypnotics.
Direct, clinically relevant specific indicators of the nociceptive–anti-nociceptive balance during general anaesthesia do not exist. Therefore, estimation of this balance is commonly based on isolated, unspecific autonomic reactions, such as the presence or absence of hypertension, tachycardia, or sweating and tearing. Another merely experimental approach, which estimates only the anti-nociceptive side of the balance, is to determine directly or to model plasma concentrations of opioids.4 Finally, an unambiguous clinical end-point, like movement in response to surgery, is usually considered as the ultimate indicator of inadequate analgesia. Unfortunately, these measures of nociception are subjected to inter-individual variability, are rather non-specific, that is, are influenced by confounding factors such as vasoactive drugs, may lack sensitivity to actual nociception and are generally without real predictive value. Hence, a multiparameter approach might improve the estimation of nociceptive–anti-nociceptive balance during general anaesthesia.5
In clinical research, several quantitative scores have been developed to classify a subject's state during general anaesthesia. Observer's Assessment of Alertness/sedation Scale (OAA/S)6 is commonly used to describe the effect of hypnotics and test the performance of ‘awareness’ monitors.7–9 The relationship between stages of ether anaesthesia and clinical signs and reflexes were first described by Guedel in 1937.10 However, standard clinical scores to evaluate the nociceptive–anti-nociceptive balance in an unconscious subject and suitable for modern anaesthetic practice do not exist. It complicates the evaluation of a potential analgesia monitor. To evaluate any objective index based on physiological parameters that aims at measuring presence and/or severity of nociception, a clinical reference is required. A usable measure of the nociceptive–anti-nociceptive balance, either clinical or paraclinical, should unambiguously discriminate patients with clinical signs of nociception from those without them. It should also reflect changes in the level of noxious stimulation: without a noxious stimulus, no signs of nociception should be present, and the intensity of the stimuli should have a graded effect on the measure. Finally, a measure of the nociceptive–anti-nociceptive balance should react to changes in the level of an analgesic drug in the presence of a standard stimulus (i.e. the higher the dose of an analgesic drug, the lower the probability that a given noxious stimulus causes actual nociception).
In an attempt to develop such clinical and paraclinical tools, we recorded physiological responses to skin incision during propofol–remifentanil anaesthesia. We combined clinical signs, remifentanil levels and an approximation of the noxious intensity of skin incision into a Clinical Signs–Stimulus–Antinociception (CSSA) score aiming at estimating the balance between nociceptive stimulation and the anti-nociceptive effect of anaesthesia at a given moment (at skin incision in this particular study). We therefore used the relationship between CSSA and several physiological parameters to combine them into an indicator of nociception (response index of nociception, RN) that can be implemented in an on-line anaesthesia monitor to provide continuous monitoring of the nociceptive–anti-nociceptive balance.
|
Patients and methods |
|---|
|
TopFootnotesAbstractIntroductionPatients and methodsResultsDiscussionAppendixReferences |
|---|
The study was approved by the institutional review board of Tampere University Hospital, and written informed consent was obtained from all patients. We enrolled 60 females, scheduled for gynaecological or breast surgery under general anaesthesia. Inclusion criteria were as follows: age 18–70 yr, ASA physical status I or II, and BMI <30 kg m–2. Exclusion criteria were as follows: known neurological disorders, history of head injury, any medication seriously affecting cardiac or central nervous system, major cardiac disorders, uncontrolled hypertension, and history of substance abuse.
All patients were premedicated with diazepam 5–10 mg orally 60 min before surgery. In the operating room an i.v. catheter was inserted into a vein of the forearm and standard monitoring [ECG, photoplethysmography (PPG), non-invasive blood pressure] was started. After degreasing of the forehead skin using isopropanol 70%, the EEG electrode strip (Entropy Sensor, GE Healthcare, Helsinki, Finland) for EEG Spectral Entropy (later Entropy) measurement was positioned as recommended by the manufacturer. Entropy is designed to monitor the brain state by revealing anaesthetic-related alterations in EEG and electromyogram (EMG) during general anaesthesia and surgery.11 Briefly: in EEG entropy monitoring high irregularity of the EEG signal is associated with a high amount of entropy. This is typically the case in an awake subject. With deepening anaesthesia the EEG becomes more regular, and its entropy content decreases. A biosignal measured from the forehead of a patient includes a significant EMG component, which is created by muscle activity. Increased EMG activity is associated with inadequate anaesthetic state.12 When collected from the forehead with surface electrodes, the spectral range of the biosignal up to 32 Hz is usually dominated by EEG, while spectral content above 32 Hz is mostly caused by EMG activity.11 To maximize the collected information, Entropy monitoring separates the EEG-dominated frequency range of 0.8–32.0 Hz, called state entropy (SE), from the information collected at 0.8–47.0 Hz, consisting of both EEG and EMG, named response entropy (RE). Entropy parameters range from 0 (EEG suppression) to 100 (awake) for RE, and from 0 to 91 for SE, where the difference between RE and SE corresponds to the EMG activity, thus giving an idea of the adequacy of anaesthesia. In surgical anaesthesia, the recommended SE range is 40–60.13 The Entropy measurement was started before the induction of anaesthesia and continued until the end of surgery.
Patients were anaesthetized using propofol and remifentanil. These drugs were administered as target-controlled infusions (TCI, Orchestra Primea®, Fresenius, France). Both infusions were started simultaneously. For propofol, the pharmacokinetic model of Schnider and colleagues14 was used, while the model of Minto and colleagues15 was utilized for adjustment of remifentanil infusion. Propofol TCI infusion was adjusted to maintain SE between 35 and 60, the target being at 50 (adequate surgical level of hypnosis).13 Propofol infusion was adjusted during surgery as necessary. Patients were randomly allocated according to computer-generated random numbers (Excel 2000, Microsoft Corp., Redmond, WA, USA) to receive remifentanil at three different TCI levels: 1, 3 or 5 ng ml–1. These levels were achieved before tracheal intubation and maintained thereafter. Patients were immobilized using a non-depolarizing neuromuscular blocking drug; rocuronium 40 mg. Additional doses of rocuronium were later given according to clinical needs. Neuromuscular blockade was monitored every 60 s with ulnar nerve train of four (TOF) stimulation and neuromuscular transmission monitoring (S/5 Anesthesia Monitor, GE Healthcare). After securing the airway by tracheal intubation, controlled normoventilation with an air–oxygen mixture (
0.35%) was started (S/5 Aespire® ventilator, Datex-Ohmeda, Inc., Madison, USA).
The exact time of skin incision was recorded on file. A trained research nurse (M.K.), who was instructed to carefully note any clinical sign (coughing, chewing or biting, other mouth movements, muscle tension, grimacing, movements of extremities, tearing) possibly related to inadequate analgesia, monitored the patient throughout surgery, until the patient regained consciousness after the end of the procedure. The surgical team was also asked to notify for any extraordinary tension or movements. During anaesthesia, the research nurse annotated any significant events on the file. She also interviewed the patients in the post-anaesthesia care unit, regarding their possible anaesthesia-associated memories.
Data acquisition and processing
Continuously monitored signals (ECG, raw PPG waveform, RE and SE) were collected using the S/5 Anesthesia Monitor (GE Healthcare; Central® and Wincollect® softwares). Intermittently monitored parameters, such as non-invasive blood pressure, were not considered, as our goal was to obtain a continuous index for nociceptive–anti-nociceptive balance.
The ECG, PPG and Entropy signals were analysed off-line in each separate individual before (60 s period) and after (120 s period) the first incision using Matlab® software (Matlab, version 6.5, Release 13, The Mathworks Inc., MA, USA). Several physiological parameters were extracted from the signals, and significant and characteristic changes associated with skin incision were noted. Parameters were used both as absolute values and as relative values. The relative values were obtained by dividing the post-incision values by their pre-incision values, except for those variables which may have values close to zero—in that case, relative values were obtained by subtracting the pre-incision value from post-incision value.
ECG and HRV
The R-waves of the ECG signal were automatically detected and visually verified and corrected afterwards when necessary. A beat-to-beat R-to-R interval (RRI) signal was constructed as a series of time differences between the successive heart beats. Heart rate variability (HRV) was quantified both in time and frequency domain, and also by Poincaré analysis, similarly to a method reported earlier.5 RRI SD (standard deviation) was computed to describe the overall HRV. Poincaré RRI SD1 quantifies the fast beat-to-beat HRV, which is mainly related to respiratory sinus arrhythmia and vagal (parasympathetic) modulation of HRV. Poincaré RRI SD2 quantifies the slower short-term components of HRV, which are modulated both by sympathetic and parasympathetic nervous systems. Their ratio (RRI SD1/SD2) describes the ratio between fast and slow short-term HRV.
PPG
The amplitude of the raw PPG signal and the location of the dicrotic notch were detected automatically, and visually verified and corrected afterwards if needed. All heart beats with PPG artifacts were excluded from the analysis. For each heart beat, PPG amplitude and vertical PPG notch position were measured, and the corresponding beat-to-beat time-series were constructed. In addition, the PPG amplitude variability was computed using Poincaré analysis and PPG SD1, SD2 and SD1/SD2 were computed, similarly to the RRI signal. Pulse transition time (PTT) was obtained as the time from the R-wave of the ECG to the maximum derivative of the PPG waveform.
CSSA score
The details of the CSSA score are provided in Table 1. Two anaesthesiologists (M.R. and A.Y.-H.) estimated retrospectively the intensity of nociceptive signs, based on the annotations documented by the research nurse, who carefully noted any clinical sign possibly related to inadequate analgesia. During this estimation process, the anaesthesiologists were blinded to any other clinical information such as haemodynamic data. The stimuli were classified according to the size of skin incision (small incision for laparoscopy or vaginal and minor breast surgery; large incision for laparotomy and major breast surgery). Remifentanil effect-site concentration (1, 3 or 5 ng ml–1) dictated the supposed impact of anti-nociception in the score. Finally, a scaling constant was applied, giving a theoretical CSSA range of 0–12. The values of the CSSA variables (clinical signs, stimulus, analgesic drug, constant) were summed to obtain the CSSA of each respective patient.
|
Statistics and development of RN
Exploratory analysis was carried out to find out which physiological parameters would predict the clinical assessment of the probability of nociception at the time of incision (CSSA score). For this purpose, differences in distributions of post- and pre-incision values of the parameters were evaluated using the Wilcoxon signed rank test. Bonferroni-correction was not used.5 Then the data were classified into groups on the basis of the remifentanil TCI level (1, 3 or 5 ng ml–1) during the incision, the reaction of each patient to skin incision (non-movers, movers), and on the basis of the size of incision (small incision for laparoscopy or vaginal and minor breast surgery, large incision for laparotomy and major breast surgery). Kruskal–Wallis H nonparametric test was applied to test the differences between the remifentanil groups, Mann–Whitney's U-test for testing between mover and non-mover groups, and between small and large incision groups. Data are presented as means (SD) unless otherwise indicated.
An algorithm using combinations of parameters was developed to map them into the three components of CSSA: level of the analgesic drug, level of stimulation, and presence of clinical signs. The outputs of these three modules were then combined into an index (RN) estimating CSSA on the basis of the monitored signals. This combination was optimally performed using linear regression. For practical reasons (e.g. in presenting it as a number in an on-line monitoring system) the estimator was scaled from 0 to 100. In this scaled version, denoted RN, a value of 0 corresponds to a CSSA value of 0 and an RN value of 100 corresponds to CSSA value 10 or higher. The algorithm is described in more detail in the Appendix.
The performance of the RN was assessed by the root mean square error (RMSE) and Pearson's correlation coefficient for the module estimating remifentanil level, and prediction probability (pk) for the components reflecting clinical observations and stimulus intensity.16 Estimations of pk and significance of differences between pk values were obtained using the Excel macro's PKMACRO and PKDMACRO as made available by Smith and colleagues.16 An estimation of the significance of the obtained value of pk was acquired by calculating the Z-score. The associated P-value indicates the probability of being wrong when asserting that the pk value is different from 0.5, which is a pk value associated to a prediction probability based on chance alone. Finally, the RN value was used in a binary logistic regression function, and the classification performance of the function was assessed by computing the area under the receiver operator curve (ROC).
Behaviour of RN in additional patients
In order to evaluate the behaviour of RN in an independent patient group, an additional 10 females (ASA physical status I or II, age 18–70 yr, BMI <30 kg m–2), scheduled for comparable procedures under propofol–remifentanil anaesthesia, were studied. Their remifentanil effect-site concentrations were recorded, but not standardized, and their data were not used in the development of RN. Research-related notes were collected as described above.
|
Results |
|---|
|
TopFootnotesAbstractIntroductionPatients and methodsResultsDiscussionAppendixReferences |
|---|
Five patients were excluded because of technical problems (n=3) or ECG left bundle branch block (n=2). The remaining 55 patients were included. Their mean age was 46 (23–66) yr and BMI 25.5 (5.6) m–2. Regarding rocuronium consumption and TOF data, no differences were detected between remifentanil groups, small or big incisions, or movers and non-movers. The group distribution of the patients and their TOF data are summarized in Table 2. No anaesthesia-related memories were reported during the postoperative interview. There were five patients who moved in response to skin incision, all in the small incision group and at a remifentanil concentration of 1 ng ml–1 (Table 2). To balance for this, the statistical comparison of the parameters between small and large incision and between different remifentanil levels was based only on subjects who did not show any clinical sign at skin incision.
|
CSSA score and basic variables
Recorded CSSA values ranged between 2 and 8 (Fig. 1).
|
Several physiological parameters differed between the pre- and post-incision periods (Table 3). The RRI decreased (i.e. heart rate increased) while the short (RRI SD1) and longer-term HRV (RRI SD, RRI SD2) increased. The PPG amplitude decreased, PPG slow variability (PPG SD2) increased and RE increased as a response to skin incision. PTT decreased.
|
Among the different remifentanil levels, between the non-movers and movers, and small vs large incision, the physiological parameters differed only marginally. Heart rate response to skin incision diminished, the response of the fast component of HRV (RRI SD1) increased, the change in RE decreased, and the change in PTT increased with higher doses of remifentanil (remifentanil 1 ng ml–1 vs 3 or 5 ng ml–1, P<0.05 in each variable). RE–SE difference increased in movers but not in non-movers. The responses of the other raw physiological parameters did not differ significantly between the different remifentanil levels, with the incision size or according to the presence/absence of clinical signs.
RN behaviour
We reached the correlation coefficient of 0.75 (P<0.001) (95% confidence interval 0.61–0.85) and RMSE 1.1 ng ml–1 for estimation of remifentanil concentration; pk of 0.87 (SE=0.06) (P<0.001) for clinical observation; and pk of 0.68 (SE=0.08) (P=0.02) for stimulus intensity. The pk of the eventual CSSA estimation in the original 55 patients was 0.78 (SE=0.05) (P<0.001), and in the additional 10 patients 0.73 (SE=0.17) (P=0.10). The difference pk (0.04, SE=0.18) was not significant (P=0.82). Figure 2 shows the behaviour of RN, plotted as a function of CSSA score.
|
To study the immediate response to skin incision RN was averaged over 20–60 s after the incision (computing delay, i.e.
20 s time to maximum taken into account). For comparison, these figures are added to Table 3. Patients with clinical signs (movers, n=5) showed higher RN than the others [49 (16) vs 36 (8), P=0.02, df (degrees of freedom)=1]. For non-movers (n=50), RN was higher after large incision than after small incision [39 (9) vs 35 (7), P=0.04, df=1]. Low doses of remifentanil were associated with higher RN values: 43 (9); 36 (6) and 32 (6), for 1, 3 and 5 ng ml–1 of remifentanil, respectively (P=0.0003, df=2) (Fig. 3). Area under ROC for RN and clinical observations (movement vs no movement) was 0.820 (P=0.019), for stimulus intensity (large vs small incision) 0.67 (P=0.053), and for remifentanil concentration (1 ng ml–1 vs >1 ng ml–1) 0.848 (P<0.001).
|
For comparison, we also calculated pk and correlation coefficients for each individual physiological parameter. The best parameters are reported. Post-incision RRI correlated with remifentanil concentration [r=0.52 (confidence interval 0.30, 0.69), P<0.001]. The pk of RE for clinical signs was 0.75 (SE=0.16), and pk of PPG SD2 for large vs small incision was 0.64 (SE=0.08).
|
Discussion |
|---|
|
TopFootnotesAbstractIntroductionPatients and methodsResultsDiscussionAppendixReferences |
|---|
We developed a multivariate modular algorithm to estimate nociception at the time of skin incision. This modular approach provides a more robust, easy maintainable and tunable, and intuitive implementation as compared with an approach in which the final index would be estimated in one step. The algorithm displays more marked response after large than small incision, differentiates the movers from non-movers, and classifies logically the patient groups receiving different effect-site concentrations of remifentanil. Thus, the algorithm seems to adequately monitor the crucial components of the nociceptive–anti-nociceptive balance, that is, the intensity of noxious stimulus and drug effect.
Furthermore, we presented a new clinical score for the assessment of severity of nociception in an anaesthetized patient. In the CSSA score, we assume that the three elements which are used to assess the probability of actual nociception at a given time instant (clinical signs, intensity of the stimulus, and level of analgesic drug), are separate additive factors, whose contributions are of equal importance. This score was used as a clinical reference in the development of RN. The development was based on the combination of several physiological parameters reflecting the stress response to potential nociception. It should be understood that the score was developed heuristically; using our clinical knowledge and judgement—we do not claim that the score would provide an unambiguous evaluation of the actual nociceptive–anti-nociceptive balance of the patient. As generally acknowledged, objective, exact clinical assessment of actual nociceptive–anti-nociceptive balance during general anaesthesia is a task of the utmost challenge. However, CSSA offers an approach as a clinical reference and as such may be evaluated and used while developing methods for objective measurement of nociception based on physiological parameters. This is supported by the fact that our RN, optimized to CSSA, successfully separated the reactions to similar stimuli at different remifentanil levels, stimuli with differing intensities, or patients exhibiting major clinical signs from the others. RN also showed improved performance over individual physiological parameters. No single physiological parameter tested in this study was exclusively associated with all modules of interest, that is, remifentanil concentration, presence of clinical signs, and intensity of surgical stimulation. Hence, calculation of RN provides significantly better indication of nociceptive–anti-nociceptive balance than the individual physiological parameters tested. Furthermore, RN behaved in a similar manner for both the original patient data (development set) and the additional patient data. This suggests that the RN may potentially be applied successfully also for other patient data than those used in this particular study.
In our data, clinical signs were shown only in patients with small incision (laparoscopy) and receiving low doses of remifentanil (1 ng ml–1). This caused some imbalance in the design and we controlled it by excluding the movers from statistical analysis, which aspired to explore the effects of incision size and remifentanil level on the haemodynamic and RN responses. This was also taken into account in the tuning of the RN module for stimulus intensity estimation. Hence, we believe this imbalance is controlled in our analysis.
The pk of RN for CSSA estimation was 0.78, which may appear as a relatively low value as compared with pk values for awareness monitors such as BIS or Entropy, which are typically well above 0.90 for OAA/S. This may be because of several reasons. First, RN is largely based on the assessment of autonomic nervous system mediated reactions, which are not specific for surgical stress and nociception. Although we aimed at making RN as specific as possible for stress reactions induced by nociception, perfect specificity is rarely achieved, resulting in a lowered pk. Second, and maybe even more importantly, our clinical reference (CSSA score) is less exactly defined as clinical references used for awareness monitors (e.g. OAA/S score or clinical endpoints). For instance, we estimated the nociceptive effect of skin incision according to the size of incision, assuming that large incision is more nociceptive than small incision, because more nociceptors are involved. Interestingly, all recorded movements associated with small incisions (laparoscopy), in patients receiving remifentanil at an effect-site concentration of 1 ng ml–1. The association of clinical signs with small incision can be merely a coincidence, still challenging the hypothesis that a small incision is less noxious than a large incision. However, several factors like the location of incision, skin sensitivity, possible inflammation, etc. also influence the intensity of noxious stimulation. These factors are difficult to estimate and impossible to control totally in clinical research. As a consequence, in further studies mutual weights and/or clinical criteria of the CSSA score might be reconsidered, especially the stimulus intensity estimation: the impact of the length of skin incision on nociceptive–anti-nociceptive balance seems in our data smaller than the influence of adequate remifentanil effect-site concentration. Furthermore, our current CSSA score as a scale is not necessarily fully monotonous. The high and low CSSA values are undoubtedly associated with high and low probability of nociception, respectively, but close values of the scale may not always be correctly ranked in each individual patient and situation (e.g. a value of 5 does not necessarily differ significantly from value 6). As a result, when this scale is used as a reference, high values for pk are beyond reach. Hence, we think that the pk value of 0.78 is acceptable. Also its estimated P-value (P<0.001), indicating clearly significant difference from a pk value of 0.5, illustrates the practical value of RN.
After introducing EEG-based indices for monitoring the hypnotic component of anaesthesia, nociceptive–anti-nociceptive balance remains the final challenge for the quantification of pharmacodynamic components of general anaesthesia. Intraoperative nociception is associated with potentially harmful events; for example, stress response and its consequences, like uncontrolled haemodynamics, increased muscle tension, and movement. On the other hand, the use of opioids is shown to increase the incidence of postoperative nausea and vomiting, a complication of general anaesthesia with significant negative impact on patients' well-being after surgery.17 Avoidance of both unnecessarily low and high doses of opioids is therefore recommended. Possibility to reliably monitor intraoperative nociceptive events and, accordingly, readjust drug delivery possesses strong clinical value that can improve the quality of anaesthetic care.
The data directly linking improved intraoperative antinociception to better postoperative outcome are sparse and confused.18 19 As no methods to objectively quantify the severity of nociception have been available, estimations of a patient's intraoperative state are based merely on given analgesics or clinical signs of nociception. Our approach offers a tool to quantify the nociception-associated factors at the patient level, instead of relying only on group statistics.
We used propofol-remifentanil anaesthesia. Propofol is prone to alter the hypnotic component of anaesthesia, here measured with Entropy. We kept the Entropy level constantly close to 50 by adjusting propofol TCI accordingly. The influence of intensity of hypnotic effect of propofol; that is ‘light’ vs ‘deep’ anaesthesia, on RN remains to be studied in another prospective investigation.
The physiological data used in the development of RN were collected in relatively healthy persons, and further tested in a limited number of comparable subjects. Performance of the model with other drug combinations remains to be tested in further prospective studies. As calculation of RN uses input variables like heart rate, HRV and PPG, dysfunction of autonomic nervous system, sometimes associated with diabetes or neuropathies, and medical conditions like cardiac arrhythmias or need of cardiac pace-maker may deteriorate the performance of the measure. Perioperative use of ß-blocking, vasoactive or parasympatolytic drugs, affecting heart rate control, may affect RN. Also pathophysiological states like hypovolaemia or hypothermia may change the behaviour of PPG and, accordingly, affect RN. Finally, as the RE–SE gradient of Entropy is sensitive to adequacy of anaesthesia, the degree of neuromuscular block can influence the gradient and, subsequently, RN, at least without adequate anti-nociceptive medication. In this study, the degree of neuromuscular block was similar in all study groups, but the theoretical possibility needs to be recognized. As the present study focused solely on a clearly specified noxious event (skin incision), further research is needed to test the feasibility of RN during surgery.
In conclusion, we developed a robust, easily maintainable and tunable modular algorithm to estimate nociception at the time of skin incision. The algorithm differentiates the movers from non-movers, displays more marked response after large than small incision, and also differentiates the patient groups receiving different concentrations of remifentanil during propofol anaesthesia. We also presented a new clinical score for the assessment of severity of nociception in an anaesthetized, paralysed patient, based on a combination of several physiological parameters. Such an approach may add value to clinical anaesthesia monitoring, but further feasibility studies are needed to show its behaviour during surgery.
|
Appendix |
|---|
|
TopFootnotesAbstractIntroductionPatients and methodsResultsDiscussionAppendixReferences |
|---|
Description of the algorithm to calculate RN
Nociception is dependent on multiple factors and is largely a nonlinear phenomenon. To better cope with it a modular algorithm with linear and nonlinear parts was designed. The algorithm uses three modules. Each of them was designed to estimate one of the components of the CSSA score. The components are presence or absence of clinical signs of nociception, stimulus intensity and antinociceptive effect. Each module uses a multivariate equation: the antinociceptive effect is estimated using a linear equation; the two other components are estimated using a nonlinear function (sigmoid function). The outputs of these three equations are then combined linearly to obtain the RN values estimating the CSSA score (Fig. 4). Nonlinear scaling was applied for linear models to better cope with nonlinearity of the phenomenon.
|
Estimation of antinociceptive effect (remi_est) is performed by a linear equation:
 | (A.1) |
Estimation of clinical signs of nociception (clinical signs_est) is implemented by a sigmoid function:
 | (A.2) |
Estimation of the stimulus intensity (stimulus_est) is implemented by another sigmoid function:
 | (A.3) |
Finally, the estimation of CSSA and scaling to RN (CSSA_est) is obtained by linear equation:
 | (A.4) |
Again, a, b, c and d are positive constants. Finally, the CSSA_est value is scaled to fit to a range of 0–100 to form the final RN value—this is done by multiplying with a factor 10 and clipping RN values to a maximum of 100.
Note: To obtain a smoother behaviour a smoothing function can be applied (e.g. in the example curves of Fig. 3, a 20-s Hanning window was used).
|
Footnotes |
|---|
|
TopFootnotesAbstractIntroductionPatients and methodsResultsDiscussionAppendixReferences |
|---|
Presented in part at the annual Euroanaesthesia meeting, Vienna, Austria, May 28–31, 2005. 
Declaration of interest. Dr Rantanen has received two direct research grants from GE Healthcare, Finland. Dr Yli-Hankala is a paid medical advisor for GE Healthcare Finland, Helsinki, Finland. Dr Huiku, Mr Takala and Mr Kymäläinen are employees of GE Healthcare, Finland. Dr Seitsonen has received equipment and personnel support from Datex-Ohmeda for her previous adequacy of anaesthesia-related studies. Both Tampere University Hospital, Department of Anaesthesia, and VTT Information Technology have received funding for this particular study from GE Healthcare, Finland.
|
References |
|---|
|
TopFootnotesAbstractIntroductionPatients and methodsResultsDiscussionAppendixReferences |
|---|
1 Yaksh TL. An introductory perspective on the study of nociception and its modulation. In Yaksh TL, Lynch C,III, Zapol WM, Maze M, Biebuyck JF and Saidman LJ (eds). Anesthesia—Biologic Foundations. Philadelphia: Lippincott–Raven Publishers, 1998; PP. 471–82
2 Sherrington CS. The Integrative Action of the Nervous System. Cambridge: The University Press, 1906
3 Kern SE, Xie G, White JL, Egan TD. Opioid-hypnotic synergy. Anesthesiology 2004; 100: 1373–81[Web of Science][Medline]
4 Boillon TW, Bruhn J, Radulescu L, et al. Pharmacodynamic interaction between propofol and remifentanil regarding hypnosis, tolerance of laryngoscopy, bispectral index, and electroencephalographic approximate entropy. Anesthesiology 2004; 100: 1353–72[CrossRef][Web of Science][Medline]
5 Seitsonen ERJ, Korhonen IKJ, van Gils MJ, et al. EEG spectral entropy, heart rate, photoplethysmography and motor responses to skin incision during sevoflurane anaesthesia. Acta Anaesthesiol Scand 2005; 49: 284–92[CrossRef][Web of Science][Medline]
6 Chernik DA, Gillings D, Laine H, et al. Validity and reliability of the Observer's Assessment of Alertness/Sedation Scale: study with intravenous midazolam. J Clin Psychopharmacol 1990; 10: 244–51[Web of Science][Medline]
7 Glass PS, Bloom M, Kearse L, et al. Bispectral analysis measures sedation and memory effects of propofol, midazolam, isoflurane, and alfentanil in healthy volunteers. Anesthesiology 1997; 86: 836–47[CrossRef][Web of Science][Medline]
8 Viertiö-Oja H, Meriläinen P, Särkelä M, et al. Spectral entropy, approximate entropy, complexity, fractal spectrum, and bispectrum of EEG during anesthesia. In: Proceedings of the 5th International Conference on Memory, Awareness, and Consciousness. New York: Memorial Sloan-Kettering Cancer Center 2001; 49–50
9 Kurita T, Doi M, Katoh T, et al. Auditory evoked potential index predicts the depth of sedation and movement in response to skin incision during sevoflurane anesthesia. Anesthesiology 2001; 95: 364–370[CrossRef][Web of Science][Medline]
10 Guedel AE. Inhalational Anesthesia: a Fundamental Guide. New York: Macmillan, 1937
11 Viertiö-Oja H, Maja V, Särkelä M, et al. Description of the Entropy algorithm as applied in the Datex-Ohmeda S/5 Entropy Module. Acta Anaesthesiol Scand (Suppl.) 2004; 48: 154–61[CrossRef][Web of Science][Medline]
12 Paloheimo M. Quantitative surface electromyography (qEMG) applications in anaesthesiology and critical care. Acta Anaesthesiol Scand 1990; 93: 1–83
13 Vakkuri A, Yli-Hankala A, Talja P, et al. Time-frequency balanced spectral entropy as a measure of anesthetic drug effect in central nervous system during sevoflurane, propofol, and thiopental anesthesia. Acta Anaesthesiol Scand 2004; 48: 145–53[CrossRef][Web of Science][Medline]
14 Schnider TW, Minto CF, Gambus PL, et al. The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers. Anesthesiology 1998; 88: 1170–82[CrossRef][Web of Science][Medline]
15 Minto CF, Schnider TW, Shafer SL. Pharmacokinetics and pharmacodynamics of remifentanil. II. Model application. Anesthesiology 1997; 86: 24–33[CrossRef][Web of Science][Medline]
16 Smith WD, Dutton RC, Ty Smith N. Measuring the performance of anesthetic depth indicators. Anesthesiology 1996; 84: 38–51[CrossRef][Web of Science][Medline]
17 Hofer CK, Zollinger A, Büchi S, et al. Patient well-being after general anaesthesia: a prospective, randomized, controlled multi-centre trial comparing intravenous and inhalation anaesthesia. Br J Anaesth 2003; 91: 631–7
18 Vinik HR, Kissin I. Rapid development of tolerance to analgesia during remifentanil infusion in humans. Anesth Analg 1998; 86: 1307–11[Abstract]
19 Lee LHY, Irwin MG, Lui SK. Intraoperative remifentanil infusion does not increase postoperative opioid consumption compared with 70% nitrous oxide. Anesthesiology 2005; 102: 398–402[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Gruenewald, P. Meybohm, C. Ilies, J. Hocker, R. Hanss, J. Scholz, and B. Bein Influence of different remifentanil concentrations on the performance of the surgical stress index to detect a standardized painful stimulus during sevoflurane anaesthesia Br. J. Anaesth., October 1, 2009; 103(4): 586 - 593. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Kallio, L. I. Lindberg, A. S. Majander, K. H. Uutela, M. L. Niskanen, and M. P. J. Paloheimo Measurement of surgical stress in anaesthetized children Br. J. Anaesth., September 1, 2008; 101(3): 383 - 389. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Weil, S. Passot, F. Servin, and V. Billard Does Spectral Entropy Reflect the Response to Intubation or Incision During Propofol-Remifentanil Anesthesia? Anesth. Analg., January 1, 2008; 106(1): 152 - 159. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Rantanen, H. Ypparila-Wolters, M. van Gils, A. Yli-Hankala, M. Huiku, M. Kymalainen, and I. Korhonen Tetanic stimulus of ulnar nerve as a predictor of heart rate response to skin incision in propofol remifentanil anaesthesia Br. J. Anaesth., October 1, 2007; 99(4): 509 - 513. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. M. R. F. Struys, C. Vanpeteghem, M. Huiku, K. Uutela, N. B. K. Blyaert, and E. P. Mortier Changes in a surgical stress index in response to standardized pain stimuli during propofol remifentanil infusion Br. J. Anaesth., September 1, 2007; 99(3): 359 - 367. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M. Mathews, P. M. Cirullo, M. M. R. F. Struys, T. De Smet, R. J. Malik, C. L. Chang, and G. G. Neuman Feasibility study for the administration of remifentanil based on the difference between response entropy and state entropy Br. J. Anaesth., June 1, 2007; 98(6): 785 - 791. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Huiku, K. Uutela, M. van Gils, I. Korhonen, M. Kymalainen, P. Merilainen, M. Paloheimo, M. Rantanen, P. Takala, H. Viertio-Oja, et al. Assessment of surgical stress during general anaesthesia Br. J. Anaesth., April 1, 2007; 98(4): 447 - 455. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Ahonen, R. Jokela, K. Uutela, and M. Huiku Surgical stress index reflects surgical stress in gynaecological laparoscopic day-case surgery Br. J. Anaesth., April 1, 2007; 98(4): 456 - 461. [Abstract] [Full Text] [PDF]
|
|
E-letters:
Read all E-letters
- Multivariable not multiparameter
- Neville W Goodman
- British Journal of Anaesthesia, 13 Mar 2006 [Full text]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||