The ability of bispectral index to detect intra‐operative wakefulness during isoflurane/air anaesthesia,compared with the isolated forearm technique

Clinical signs are unreliable for guiding anaesthetic administration and it is suggested that using the bispectral index can improve anaesthetic delivery. In the current study, isoflurane administration was guided to a bispectral index range of 55–60. Intra‐operative responsiveness, as assessed by the isolated forearm technique, was compared with whether the bispectral index predicted/identified a patient's appropriate hand movements in response to commands. Thirty‐four women underwent major gynaecological surgery with isoflurane/air and atracurium. Eleven women responded on 32 occasions with appropriate hand movements to commands given during surgery, of which the bispectral index detected 17 (sensitivity 53%). The bispectral index suggested consciousness 660 times in the absence of any movement responses (specificity 69%). The positive predictive value of the bispectral index was 3%. The median (IQR [range]) bispectral index value associated with an intra‐operative response was significantly lower than that associated with eye opening after surgery: 60 (50–68 [36–83]) vs 77 (75–84 [59–90]), respectively (p = 2.25 × 10^?8). Conversely, end‐tidal isoflurane concentration was significantly higher at intra‐operative response than at eye opening: 0.3 (0.3–0.4 [0.2–0.9]) vs 0.2 (0.1–0.2 [0.1–0.3]), respectively (p = 7.36 × 10^?8). For patients who responded more than once during surgery, the bispectral index value associated with a response was not constant. No patient had recall for surgery or the taped commands, and only one could remember dreaming (a good dream). Titrating isoflurane to target a bispectral index range of 55–60 may result in an unacceptable number of patients who are conscious during surgery (albeit without recall).     

The combined effects of midazolam and propofol sedation on muscle power

We performed a randomised, crossover study to investigate the effects of intravenous sedation on grip strength and bite force. Twenty male volunteers received a bolus intravenous injection of midazolam (0.02 mg.kg^?1) together with a 30‐min propofol infusion designed to achieve an effect‐site concentration of 1.0 μg.ml^?1. Observed variables included bispectral index, observer's assessment of alertness/sedation, correct answer rate of Stroop colour‐word test, grip strength and bite force. Grip strength decreased from a median (IQR [range]) of 483 (443–517 [380–586]) N to 358 (280–405 [108–580]) N (p < 0.001) during sedation and recovered following flumazenil administration, while bite force increased from 818 (593–1026 [405–1406]) N to 1377 (1243–1585 [836–2357]) N (p < 0.001) during sedation. Although bite force gradually returned to baseline following flumazenil administration, it remained increased throughout the experimental period. We conclude that bite force increased during intravenous sedation and that this may have clinical implications.     

Reversal of neuromuscular blockade with sugammadex during continuous administration of anaesthetic agents: a double-blind randomised crossover study using the bispectral index

Sugammadex, a specific reversal agent for steroidal neuromuscular blocking drugs, has on occasion been reported to be associated with clinical signs of awakening. We performed a study to systematically search for an increase in bispectral index values and signs of awakening in patients maintained under general anaesthesia following sugammadex administration. Patients, scheduled to receive general anaesthesia with neuromuscular blockade, were included in this double-blind randomised crossover study. After surgery was completed, and while the train-of-four ratio was zero, intravenous anaesthesia was continued with the aim of maintaining the bispectral index in the range of 40–60. Patients then received either sugammadex 4 mg.kg⁻¹ or saline. In cases of incomplete reversal of neuromuscular blockade after 5 min, patients received the other drug. Bispectral index and train-of-four monitoring were recorded every minute and clinical signs of awakening noted. Fifty-one patients completed the study. Median (IQR [range]) bispectral index values increased after sugammadex administration from 49 (43–53 [38–64]) to 63 (53–80 [45–97]) (p < 0.01) with an increase of ≥ 20 in 22 patients; 14 (27%) patients had clinical signs of awakening. Saline had no effect on bispectral index values, clinical signs of awakening or degree of neuromuscular blockade. This study confirms that reversal of neuromuscular blockade with sugammadex may be associated with clinical signs of awakening despite maintenance of anaesthesia. Intravenous anaesthesia should be maintained until complete recovery of muscle function is achieved, especially when sugammadex is administered.     

The effect of remifentanil on propofol requirements to achieve loss of response to command vs. loss of response to pain

When providing total intravenous anaesthesia, careful selection of end‐points is required in titrating dose to effect during induction. Although propofol and remifentanil have predominantly different pharmacodynamic effects, they are seen to interact in achieving loss of consciousness and analgesia. To highlight these differences, we performed a double‐blind, randomised controlled trial, comparing one group of patients receiving propofol alone (n = 42) with another group receiving remifentanil plus propofol (n = 46) as a target‐controlled infusion of remifentanil (Minto; 3 ng.ml^?1). Propofol was also titrated using a target‐controlled infusion (Marsh effect model) to produce loss of response to tactile and vocal stimuli, and subsequently to loss of response to pain. The effect‐site concentration of propofol at which 50% of patients lost tactile/verbal response was 2.9 μg.ml^?1 in the propofol only group and 2.4 μg.ml^?1 in the remifentanil with propofol group. In contrast, loss of pain response occurred at 4.4 μg.ml^?1 in the propofol group, and 2.7 μg.ml^?1 in the remifentanil with propofol group, with correspondingly lower bispectral index values. Judicious use of analgesia in total intravenous anaesthesia can have a propofol‐sparing effect and potentially minimise the suppression of brain electrical activity.     

Effect of sufentanil on bispectral index in the elderly

We examined the impact of adding sufentanil during anaesthesia induction with propofol on bispectral index values in elderly patients (≥ 65 years). Patients were randomly assigned to receive a target‐controlled sufentanil infusion (effect‐site concentration of 0.3 ng.ml^?1) or matching placebo, followed by a target‐controlled propofol induction (initial effect‐site concentration of 0.5 μg.ml^?1; step‐wise increase of 0.5 μg.ml^?1) until loss of consciousness defined as an Observer's Assessment of Alertness/Sedation score < 2. Seventy‐one patients (sufentanil 35, placebo 36) completed the study. Mean (SD) age was 72.3 (5.8) years; 41% were women. At loss of consciousness, mean (SD) bispectral index value was 75.0 (8.6) with sufentanil and 70.0 (8.0) with placebo; mean difference ?5.0 (95% confidence interval ?8.9 to ?1.1), p = 0.013. Post‐hoc analyses suggest that the difference was significant in men only (mean difference ?7.3 (?11.8 to ?2.6), p = 0.003). Sufentanil co‐induction with propofol results in higher bispectral index values at loss of consciousness in elderly patients.     

The effect of propofol and desflurane anaesthesia on human hepatic blood flow: a pilot study*

This study tested the hypothesis that propofol is associated with a higher hepatic blood flow in humans compared with desflurane. Using a cross over study design, 10 patients received first propofol and then desflurane, and a further 10 patients received desflurane and then propofol. Blood flow index in the right and middle hepatic veins, stroke volume index and cardiac index were assessed by transoesophageal echocardiography. Mean arterial blood pressure, stroke volume index and cardiac index were the same in both groups. Propofol was associated with significantly greater blood flow index in the right hepatic vein (median (IQR [range]) 199 (146–237 [66–388]) vs. 149 (112–189 [42–309]) ml.min^?1.m^?2; p = 0.005) and middle hepatic vein (150 (122–191 [57–341]) vs. 125 (92–149 [47–362]) ml.min^?1.m^?2; p < 0.001) compared with desflurane. In routine clinical conditions, propofol anaesthesia was associated with significantly greater hepatic blood flow than desflurane anaesthesia.     

Absence of memory for intra-operative information during surgery with total intravenous anaesthesia

While using the isolated forearm technique, we wished to determinewhether patients who did not respond to commands during generalanaesthesia with a total intravenous technique (propofol andalfentanil with atracurium) had any evidence of post-operativeexplicit or implicit memory. Forty women undergoing major gynaecologicalsurgery were randomized, in a double-blind design, to hear twodifferent tapes during surgery. Psychological tests of explicitand implicit memory were conducted within 2 h of surgery. Therewas no evidence of implicit or explicit memory, nor any recall,in the seven women who responded to commands during surgery.We conclude that during total intravenous anaesthesia with propofoland alfentanil, there is no evidence that learning takes placewhen anaesthesia is adequate. Furthermore, with this anaesthetictechnique, it would seem that—provided any period of patientresponsiveness is short and that unconsciousness is inducedrapidly again—there is no evidence of implicit or explicitmemory. Br J Anaesth 2001; 86: 196–202     

Clinical predictors of apnoea-hypopnoea during propofol sedation in patients undergoing spinal anaesthesia

This study assessed the relationship between the occurrence of apnoea-hypopnoea during propofol sedation for spinal anaesthesia and two different predictive tests of sleep apnoea: the STOP-Bang score (snoring while sleeping, daytime tiredness, observed breathing stoppages, high blood pressure-body mass index, age, neck circumference, gender); and the obstructive sleep apnoea (OSA) score. Thirty-four middle-aged men not diagnosed with obstructive sleep apnoea received propofol infusions adjusted to produce a bispectral index of 70-75. ApnoeaLink(TM) was used to estimate the incidence of apnoea-hypopnoea. The median (IQR [range]) apnoea-hypopnoea index was 17 (8-24 [0-70]) events.h(-1) and correlated weakly with the STOP-Bang score (p = 0.022, r = 0.423) and moderately with the OSA score (p < 0.001, r = 0.693). Severe apnoea-hypopnoea developed more frequently in patients with a higher OSA score (34.5% vs 0%) or higher STOP-Bang score (27.6% vs 6.9%). Both assessment tools have some predictive value for the occurrence of apnoea-hypopnoea during propofol sedation in patients undergoing spinal anaesthesia.     

ORIGINAL ARTICLE: Optimal effect‐site concentration of remifentanil for preventing cough during emergence from sevoflurane‐remifentanil anaesthesia*

This randomised, double‐blinded, controlled trial was designed to identify the optimal dose of remifentanil for cough suppression without adverse effects during emergence from sevoflurane‐remifentanil anaesthesia for thyroidectomy. One hundred and four patients were randomly assigned to maintain target effect‐site concentrations of remifentanil at 0 (control group), 1.0 (remifentail 1 group), or 1.5 ng.ml^?1 (remifentanil 1.5 group) during emergence. The incidence of coughing was lower in the remifentanil 1.5 group (31%) than in the control group (74%) or remifentanil 1 group (63%) (p = 0.0004). In addition, the severity of coughing during extubation was lower in the remifentanil 1.5 group (median (IQR [range]) 0 (0–1 [0–1]) than in the control group (1 (0–2 [0–3])) and remifentanil 1 group (1 (0–2 [0–3])) (p = 0.004). Haemodynamic changes were reduced, but emergence time and stay in the post‐anaesthesia care unit was prolonged in the remifentanil 1.5 group. Maintaining the remifentanil effect‐site concentration at 1.5 ng.ml^?1 during emergence from sevoflurane‐remifentanil anaesthesia reduces the incidence and severity of coughing without serious adverse events and may provide haemodynamic stability in patients undergoing thyroidectomy. However, awakening may be delayed.     

Postoperative morphine concentration in infants with or without biliary atresia and its association with hepatic blood flow

We measured pre‐operative hepatic blood flow and postoperative morphine concentration in infants with or without biliary atresia. Thirty‐four infants (0–3 months) with biliary atresia undergoing portoenterostomy (Kasai operation) were included and hepatic blood flow was assessed by magnetic resonance imaging before surgery in 12 of them. Sixteen subjects (0–3 months) without liver disease undergoing abdominal or pelvic surgery acted as controls and six of them had hepatic blood flow assessed. Intravenous morphine (8 μg.kg^?1.h^?1) was administered to all patients postoperatively. The median (IQR [range]) relative hepatic blood flow was 3.51 (2.72–3.88 [1.68–4.43]) with and 3.15 (2.66–4.42 [2.30–5.01]) without biliary atresia (p = 0.851). The median (IQR [range]) morphine concentration after 24 h infusion was 5.9 (4.5–16.4 [2.9–42.2]) ng.ml^?1 and 6.4 (3.2–12.0 [1.9–48.6]) ng.ml^?1, respectively (p = 0.460). An inverse regression relation was found between the morphine concentration and the hepatic perfusion index (R² = 0.519, p = 0.001). Compensatory increases in hepatic arterial blood flow maintain the total hepatic blood flow in infants with biliary atresia.     

Transnasal Humidified Rapid‐Insufflation Ventilatory Exchange (THRIVE): a physiological method of increasing apnoea time in patients with difficult airways

Emergency and difficult tracheal intubations are hazardous undertakings where successive laryngoscopy–hypoxaemia–re‐oxygenation cycles can escalate to airway loss and the ‘can't intubate, can't ventilate’ scenario. Between 2013 and 2014, we extended the apnoea times of 25 patients with difficult airways who were undergoing general anaesthesia for hypopharyngeal or laryngotracheal surgery. This was achieved through continuous delivery of transnasal high‐flow humidified oxygen, initially to provide pre‐oxygenation, and continuing as post‐oxygenation during intravenous induction of anaesthesia and neuromuscular blockade until a definitive airway was secured. Apnoea time commenced at administration of neuromuscular blockade and ended with commencement of jet ventilation, positive‐pressure ventilation or recommencement of spontaneous ventilation. During this time, upper airway patency was maintained with jaw‐thrust. Transnasal Humidified Rapid‐Insufflation Ventilatory Exchange (THRIVE) was used in 15 males and 10 females. Mean (SD [range]) age at treatment was 49 (15 [25–81]) years. The median (IQR [range]) Mallampati grade was 3 (2–3 [2–4]) and direct laryngoscopy grade was 3 (3–3 [2–4]). There were 12 obese patients and nine patients were stridulous. The median (IQR [range]) apnoea time was 14 (9–19 [5–65]) min. No patient experienced arterial desaturation < 90%. Mean (SD [range]) post‐apnoea end‐tidal (and in four patients, arterial) carbon dioxide level was 7.8 (2.4 [4.9–15.3]) kPa. The rate of increase in end‐tidal carbon dioxide was 0.15 kPa.min^?1. We conclude that THRIVE combines the benefits of ‘classical’ apnoeic oxygenation with continuous positive airway pressure and gaseous exchange through flow‐dependent deadspace flushing. It has the potential to transform the practice of anaesthesia by changing the nature of securing a definitive airway in emergency and difficult intubations from a pressured stop–start process to a smooth and unhurried undertaking.     

A randomised clinical trial comparing the flexible fibrescope and the Pentax Airway Scope (AWS)® for awake oral tracheal intubation

We compared awake fibreoptic intubation with awake intubation using the Pentax Airway Scope^® in 40 adult patients. Sedation was achieved using a target‐controlled remifentanil infusion of 1–5 ng.ml^?1 and midazolam. The airway was anaesthetised with lidocaine spray and gargle. The total procedure time – a composite of sedation time, topical anaesthesia time and intubation time – was recorded. The operator's impression of the ease of the procedure and the patients' reported comfort were recorded on a 0–100 mm visual analogue scale. The median (IQR [range]) for total procedure time was 900 (739–1059 [616–1215]) s with the fibrescope and 651 (601–720 [498–900]) s with the Pentax Airway Scope (p = 0.0001). The median (IQR [range]) intubation time was 420 (283–480 [120–608]) s with the fibrescope and 183 (144–220 [107–420]) s with the Pentax Airway Scope (p = 0.0002). The median (IQR [range]) visual analogue scores for the operator's ease of intubation for the fibrescope and Pentax Airway Scope were 83.6 (72.0–98.0 [49.0–100.0]) and 86.8 (84.0–91.0 [61.0–100.0]), respectively (p = 0.3507). The median (IQR [range]) visual analogue score for patient comfort was 85.5 (81.0–97.0 [69.0–100.0]) and 79.4 (74.0–85.0 [59.0–100.0]) for the fibrescope and Pentax Airway Scope, respectively (p = 0.06). Total procedure time was significantly shorter with the Pentax Airway Scope compared with the fibrescope, with no difference in procedure difficulty or patient discomfort.     

A randomised controlled trial of intrathecal blockade versus peripheral nerve blockade for day‐case knee arthroscopy

We allocated 100 patients scheduled for day‐case knee arthroscopy to unilateral spinal anaesthesia with 40 mg intrathecal hyperbaric prilocaine or to ultrasound‐guided femoral‐sciatic nerve blockade with 25 ml mepivacaine 2%, 50 participants each. The median (IQR [range]) time to walk was 285 (240–330 [160–515]) min after intrathecal anaesthesia vs 328 (280–362 [150–435]) min after peripheral nerve blockade, p = 0.007. The median (IQR [range]) time to home discharge was 310 (260–350 [160–520]) min after intrathecal anaesthesia vs 335 (290–395 [190–440]) min after peripheral nerve blockade, p = 0.016. There was no difference in time from anaesthetic preparation to readiness for surgery.     

Rapid sequence spinal anaesthesia for category‐1 urgency caesarean section: a case series

General anaesthesia is the fastest method for anaesthetising a category‐1 caesarean section but is associated with increased maternal morbidity and mortality. We describe the ‘rapid sequence spinal’ to minimise anaesthetic time. This consists of a no‐touch spinal technique, consideration of omission of the spinal opioid, limiting spinal attempts, allowing the start of surgery before full establishment of the spinal block, and being prepared for conversion to general anaesthesia if there are delays or problems. We present a case series of 25 rapid sequence spinal anaesthetics for category‐1 caesarean section. The mean (SD [range]) decision‐delivery interval was 23 (6 [14–41]) min. After excluding cases where there was an identified delay, the median (IQR [range]) time to prepare and perform the spinal was 2 (2–3 [1–7]) min, and time to develop a ‘satisfactory’ block was 4 (3–5 [2–7]) min. The total time to induce spinal anaesthesia was 8 (7–8 [6–8]) min. There were three pre‐operative conversions to general anaesthesia and three women had pain during surgery that did not require treatment. Our data indicate that one might expect to establish anaesthesia in 6–8 min using a rapid sequence spinal. Careful case selection is crucial. While rapid anaesthesia is important, the reduction of the decision‐delivery interval also requires attention to other stages in the pre‐operative process.     

Differences between bispectral index and spectral entropy during xenon anaesthesia: a comparison with propofol anaesthesia

We enrolled 114 patients, aged 65–83 years, undergoing elective surgery (duration > 2h) into a randomised, controlled study to evaluate the performance of bispectral index and spectral entropy for monitoring depth of xenon versus propofol anaesthesia. In the propofol group, bispectral index and state entropy values were comparable. In the xenon group, bispectral index values resembled those in the propofol group, but spectral entropy levels were significantly lower. Mean arterial blood pressure was higher and heart rate was lower in the xenon group than in the propofol group. Bispectral index and spectral entropy considerably diverged during xenon but not during propofol anaesthesia. We therefore conclude that these measures are not interchangeable for the assessment of depth of hypnosis and that bispectral index is likely to reflect actual depth of anaesthesia more precisely compared with spectral entropy.     

Hypnotic endpoints vs. the bispectral index, 95% spectral edge frequency and median frequency during propofol infusion with or without fentanyl

Hypnotic endpoints and/or EEG variables, e.g. bispectral index, 95% spectral edge frequency and median frequency, have been studied to monitor anaesthetic (hypnotic) depth during total intravenous anaesthesia. In this study, the relation between the hypnotic endpoints of unresponsiveness to verbal commands, loss of eyelash reflex and body movement response to mechanical nasal membrane stimulation vs. bispectral index, 95% spectral edge frequency and median frequency during propofol anaesthesia with or without fentanyl is presented. Forty-two patients were randomly assigned to receive either propofol infusion, 30 mg kg-1 h-1 (n = 22), or propofol infusion, 30 mg kg-1 h-1 + fentanyl bolus, 2 micrograms kg-1 i.v. (n = 20). Bispectral index, 95% spectral edge frequency and median frequency and propofol doses were monitored and recorded at unresponsiveness to verbal commands, loss of eyelash reflex and inhibition of nasal body movement response. The bispectral index values were significantly higher in the propofol + fentanyl compared with the propofol group, i.e. 74.7 +/- 10.9, 73.1 +/- 10.5 and 47.1 +/- 9.2 vs. 65.8 +/- 9.8, 59.6 +/- 10 and 33.8 +/- 5.7 at unresponsiveness to verbal commands, loss of eyelash reflex and inhibition of nasal body movement response respectively. Doses of propofol for achieving the hypnotic endpoints were significantly lower in the propofol + fentanyl compared with the propofol group. Plasma propofol concentrations at inhibition of nasal body movement response were lower in the propofol + fentanyl compared with the propofol group (9.2 +/- 2.0 micrograms mL-1 vs. 14.1 +/- 4.2 micrograms mL-1). Our results suggest that fentanyl pretreatment potentiates the effects of propofol and achieves the hypnotic endpoints at higher bispectral index values and lower propofol doses and concentrations (measured at inhibition of nasal body movement response).     

Apnoeic oxygenation with high‐flow nasal oxygen for laryngeal surgery: a case series

Surgery under apnoeic conditions with the use of high‐flow nasal oxygen is novel. Between November 2016 and May 2017, 28 patients underwent tubeless laryngeal or tracheal surgery under apnoeic conditions with high‐flow nasal oxygen as the sole method of gas exchange. Patients received total intravenous anaesthesia and neuromuscular blocking agents for the duration of their surgery. The median (IQR [range]) apnoea time was 19 (15–24 [9–37]) min. Four patients experienced an episode of oxygen desaturation to a value between 85% and 90%, lasting less than 2 min in each case. Median (IQR [range]) end‐tidal carbon dioxide (ETCO₂) level following apnoea was 8.2 (7.2–9.4 [5.8–11.8]) kPa. The mean (SD) rate of ETCO₂ increase was 0.17 (0.07) kPa.min^?1 from an approximated baseline value of 5.00 kPa. Venous blood sampling from 19 patients demonstrated a mean (SD) partial pressure of carbon dioxide (P_VCO₂) of 6.29 (0.71) kPa at baseline and 9.44 (1.12) kPa after 15 min of apnoea. This equates to a mean (SD) P_VCO₂ rise of 0.21 (0.08) kPa.min^?1 during this period. Mean (SD) pH was 7.40 (0.03) at baseline and 7.23 (0.04) after 15 min of apnoea. Mean (SD) standard bicarbonate was 26.7 (1.8) mmol.l^?1 at baseline and 25.4 (1.8) mmol.l^?1 at 15 min. We conclude that high‐flow nasal oxygen under apnoeic conditions can provide satisfactory gas exchange in order to allow tubeless anaesthesia for laryngeal surgery.     

Changes in qualitative and quantitative ultrasound assessment of the gastric antrum before and after elective caesarean section in term pregnant women: a prospective cohort study

Ultrasound measurement of the antral cross‐sectional area allows a quantitative estimate of gastric contents in non‐pregnant adults, but this relationship may be affected by compression of the stomach exerted by the gravid uterus during pregnancy. This study aimed to assess differences in quantitative (Perlas score) and qualitative (antral cross‐sectional area) ultrasound assessments of the gastric antrum performed immediately before and after caesarean section. Forty‐three women having elective caesarean section performed under spinal anaesthesia were studied in the semirecumbent and semirecumbent‐right lateral positions. Thirty‐nine women showed no change in stomach contents using the Perlas score between the two measurement periods; four women showed a change, but by one grade only. The median (IQR [range]) antral cross‐sectional area was 323 (243–495 [103–908]) mm² before, and 237 (165–377 [112–762]) mm² after, caesarean section in the semirecumbent position (p = 0.001); the comparable values in the semirecumbent‐right lateral position were 418 (310–640 [161–1238]) mm² and 362 (280–491 [137–1231]) mm² (p = 0.09). The distance between the skin and the antrum, and the aorta and the antrum, decreased significantly in both positions after surgery. We suggest that our results indicate that stomach contents remain largely unchanged in women having elective caesarean section, but antral cross‐sectional area decreases, especially in the semirecumbent position, related to a change in the position of the stomach within the abdomen. This implies that the relationship of antral cross‐sectional area to volume of stomach contents, which has been determined for non‐pregnant subjects, may not apply in term pregnant women.     

Pulse pressure variation to predict fluid responsiveness in spontaneously breathing patients: tidal vs forced inspiratory breathing

We evaluated whether pulse pressure variation can predict fluid responsiveness in spontaneously breathing patients. Fifty‐nine elective thoracic surgical patients were studied before induction of general anaesthesia. After volume expansion with hydroxyethyl starch 6 ml.kg^?1, patients were defined as responders by a ≥ 15% increase in the cardiac index. Haemodynamic variables were measured before and after volume expansion and pulse pressure variations were calculated during tidal breathing and during forced inspiratory breathing. Median (IQR [range]) pulse pressure variation during forced inspiratory breathing was significantly higher in responders (n = 29) than in non‐responders (n = 30) before volume expansion (18.2 (IQR 14.7–18.2 [9.3–31.3])% vs 10.1 (IQR 8.3–12.6 [4.8–21.1])%, respectively, p < 0.001). The receiver‐operating characteristic curve revealed that pulse pressure variation during forced inspiratory breathing could predict fluid responsiveness (area under the curve 0.910, p < 0.0001). Pulse pressure variation measured during forced inspiratory breathing can be used to guide fluid management in spontaneously breathing patients.     

Measuring depth of anaesthesia using changes in directional connectivity: a comparison with auditory middle latency response and estimated bispectral index during propofol anaesthesia

General anaesthesia is associated with changes in connectivity between different regions of the brain, the assessment of which has the potential to provide a novel marker of anaesthetic effect. We propose an index that quantifies the strength and direction of information flow in electroencephalographic signals collected across the scalp, assess its performance in discriminating ‘wakefulness’ from ‘anaesthesia’, and compare it with estimated bispectral index and the auditory middle latency response. We used a step-wise slow induction of anaesthesia in 10 patients to assess graded changes in electroencephalographic directional connectivity at propofol effect-site concentrations of 2 μg.ml⁻¹, 3 μg.ml⁻¹ and 4 μg.ml⁻¹. For each stable effect-site concentration, connectivity was estimated from multichannel electroencephalograms using directed coherence, together with middle latency response and estimated bispectral index. We used a linear support vector machine classifier to compare the performance of the different electroencephalographic features in discriminating wakefulness from anaesthesia. We found a significant reduction in the strength of long-range connectivity (interelectrode distance > 10 cm) (p < 0.008), and a reversal of information flow from markedly postero-frontal to fronto-posterior (p < 0.006) between wakefulness and a propofol effect-site concentration of 2 μg.ml⁻¹. This then remained relatively constant as effect-site concentration increased, consistent with a step change in directed coherence with anaesthesia. This contrasted with the gradual change with increasing anaesthetic dose observed for estimated bispectral index and middle latency response. Directed coherence performed best in discriminating wakefulness from anaesthesia with an accuracy of 95%, indicating the potential of this new method (on its own or combined with others) for monitoring adequacy of anaesthesia.