Anaesthesia and Intensive Care

Occasionally accidents and complications occur during anaesthesia and perioperative care that result in injury to the patient. Unfortunately, this is sometimes due to a breach in the anaesthetist's duty of care to the patient. Sometimes, rather than being the cause of immediate damage, the act or omission results in an alteration in the prognosis of the complaint or increased risk of complications related to the complaint. This avenue for a negligence action is known as ‘loss of chance of a better outcome’ and has been the subject of much legal argument in Australia in recent years. A recent High Court of Australia decision is widely seen as having ‘closed the door’ to, or at least made it difficult for the patient to succeed in, loss of chance cases. Many anaesthetists may not be familiar with the concept of ‘loss of chance’. This review will explore the concept of loss of chance and the manner in which Australian courts have dealt with it before and after Tabet v Gett from the perspective of the anaesthetist.

Cardiac surgery is increasingly performed on elderly patients with multiple comorbid conditions, but the determinants of the relationship between cost and survival time after cardiac surgery for patients with a serious cardiac condition remain uncertain. Using the long-term outcome data of a cohort study on adult cardiac surgical patients, the relationship between cost and survival time after cardiac surgery from a hospital service perspective was determined. The total cost for each patient was estimated by the costs of the surgical procedures, intra-aortic balloon pump utilisation, operating theatre utilisation, blood products, intensive care unit stay and cumulative hospital stay up to a median follow-up time of 30 months. Of the 2131 patients considered in this study, a total cost >A$100,000 per life-year after cardiac surgery was observed only in 171 patients (8.0%, 95% confidence interval 6.9 to 9.3%). Age, Charlson Comorbidity Index and EuroSCORE were all related to the cost per life-year after cardiac surgery, but EuroSCORE (odds ratio 1.26 per score increment, 95% confidence interval 1.18 to 1.35, P=0.001) was, by far, the most important determinant and explained 32% of the variability in cost per life-year after cardiac surgery. Patients with a high EuroSCORE were associated with a substantially longer length of intensive care unit stay and cumulative hospital stay, as well as a shorter survival time after cardiac surgery compared to patients with a lower EuroSCORE. Of all the subgroups of patients examined, only patients with a EuroSCORE >5 were consistently associated with a cost >A$100,000 per life-year (cost per life-year $183,148, 95% confidence interval 125, 394 to 240, 902).

The use of cerebral oximetry to guide intraoperative management has been shown to improve patient outcomes in cardiac surgery. This pilot trial assessed the feasibility of performing a similar study of outcome in patients over the age of 70 years undergoing non-cardiac surgery. Patients over the age of 70 years undergoing total knee or hip arthroplasty or bowel resection surgery were randomly assigned to have cerebral oximetry values monitored (intervention group) or not monitored (control) while under general anaesthesia. Indicators of proof of concept were: rate of complications, rate of cerebral desaturation, relationship between cerebral desaturation and complications, and anaesthetist response to cerebral desaturation. Forty patients were recruited and randomised to a control group (n=20) or an intervention group (n=20). The proportion of the study population who had a complication was 40% (95% confidence interval [CI] 26% to 55%). Cerebral desaturation (<75% of baseline) occurred in only two patients (5.0% (CI 1.4% to 16%)), one in each group, and neither of those patients recorded a complication. Changes to anaesthetic management on the basis of cerebral oximetry values occurred in only two patients in the intervention group (10% (CI 2.8% to 30%)). Maintenance of cerebral oximetry values appeared to be closer to baseline in the intervention group than in the control group but this difference was not significant (P=0.15). Our results indicated that complications occurred frequently in the study population but did not appear to be associated with cerebral desaturation events. These findings do not support a larger intervention study using the current study population.

Children undergoing anaesthesia are prone to hypothermia. Perioperative monitoring of patient temperature is, therefore, standard practice. Postoperative temperature is regarded as a key anaesthetic performance indicator in Australian hospitals. Many different methods and sites of temperature measurement are used perioperatively. It is unclear to what degree these methods might be interchangeable. The aim of this study was to determine the relationships between temperatures measured at different sites in anaesthetised children. Two hundred children, 0 to 17 years, undergoing general anaesthesia for elective non-cardiac surgery, were prospectively recruited. Temperature measurements were taken in the operating theatre concurrently at the nasopharynx, tympanic membranes, temporal artery, axilla and skin (chest). Patient age and weight were documented. Temperatures varied according to site of measurement. The mean difference from nasopharyngeal temperature to temperatures at left and right tympanic, temporal, axillary and cutaneous sites were +0.24°C, +0.24°C, +0.35°C, −0.38°C and −1.70°C, respectively. Levels of agreement to nasopharyngeal temperature were similar at tympanic, temporal and axillary sites. Tympanic and temporal temperatures were superior to axillary temperatures for detection of mild hypothermia (<36°C). Skin temperature showed a large variation from nasopharyngeal measurements. Our findings indicate that measured temperatures vary between sites. Understanding these variations is important for interpreting temperature readings.

Assessing intravascular volume status in the critically ill patient remains a challenge for intensivists, and the accuracy of such estimation based on bedside examination alone is reported to be nearly a coin toss. In this retrospective study we sought to validate a previously recommended chest radiographic vascular pedicle width (VPW) ≥70 mm for identifying cardiogenic pulmonary oedema (CPO). We additionally assessed whether novice physicians-in-training can reliably measure the VPW. The study included intensive care patients with an existing pulmonary artery catheter. Three independent raters performed measurements of VPW from chest radiographs obtained within three hours of pulmonary artery occlusion pressure measurements. In 80 patients enrolled, a VPW cut-off of ≥70 mm had a 55% sensitivity, 88% specificity, 81% positive predictive value, 69% negative predictive value and 73% accuracy for identifying patients with CPO. Receiver operating characteristic curve analysis showed an area under the curve of 0.72 (95% confidence interval 0.61 to 0.84) for VPW in discriminating CPO from non-cardiogenic pulmonary oedema. Kappa statistics for inter-rater reliability showed Kappa=0.41, 0.42 and 0.85 for each pair of the three raters. In conclusion, the previously accepted VPW cut-off of ≥70 mm is reasonably accurate in discriminating CPO from non-cardiogenic pulmonary oedema. VPW can be measured by physicians-in-training with a comparable performance to previous studies utilising expert radiologists.

If tracheal intubation is not possible using direct laryngoscopy, one option is to use a laryngeal mask airway (LMA) through which an endotracheal tube (ETT) can be passed. In children, however, the size of an uncuffed ETT that can pass through the lumen of an LMA is sometimes too small for the trachea, resulting in gas leakage around the ETT. Using a cuffed ETT may reduce the gas leak but withdrawal of the LMA is then prevented by the pilot balloon. In this study, the largest sizes of cuffed and uncuffed Mallinckrodt™ ETTs that could pass with ease through various sizes of paediatric Classic™ and ProSeal™ LMAs were documented. For cuffed ETTs, withdrawal of the LMA was made possible by simply cutting off the pilot balloon. The ETT cuff-inflating mechanism was then repaired by passing a 20 or 22 gauge cannula into the cut end of the inflating tubing. The proximal end of the cannula was then connected to a one-way valve or a three-way stopcock. This technique of cutting off the pilot balloon of the cuffed ETT made it possible to use paediatric cuffed ETTs in exchange for the LMAs tested. The task was easy to perform. Subsequent repair of the cuff-inflation tubing was effective and could withstand high pressures. These findings indicate that it is possible to pass cuffed ETTs through paediatric LMA lumens, which can provide ventilation without gas leaks, unlike uncuffed ETTs.

Children undergoing magnetic resonance imaging (MRI) often require general anaesthesia (GA). Children under GA are at risk of decreases in body temperature. This risk may be greater during MRI due to MRI scanners requiring cool ambient temperatures. Conversely, radiofrequency radiation emitted by MRI scanners is absorbed by the patient as heat energy, creating a risk of an increase in body temperature. The aim of this study was to determine the proportion of anaesthetised children undergoing an MRI scan who develop hyperthermia or hypothermia, and the risk factors associated with temperature changes in these children. Pre-scan and post-scan tympanic temperatures were obtained from 193 children (aged three months to six years) undergoing an MRI procedure under GA. No active warming or cooling devices were used during the MRI scans. The median duration for anaesthesia was 42 minutes (35 to 57 minutes). Fifty-two percent of children were hypothermic after their scan, while no subjects were hyperthermic after their scan. The mean (± standard deviation) pre-scan temperature was 36.2°C±0.5°C, and the mean (± standard deviation) post-scan temperature was 35.9°C±0.6°C (an overall mean temperature decrease of 0.28°C was observed [95% confidence interval, −0.36°C to −0.19°C], P <0.001). In conclusion, core body temperature was found to decrease slightly during an MRI scan under GA. These results suggest that more focus is needed regarding the cooling effects of GA agents during MRI, as opposed to the heating effects of the MRI scanner.

Advanced haemodynamic monitoring employing minimally invasive cardiac output measurement may lead to significant improvements in patient outcomes in major surgery. However, the precision (scatter) of measurement of available generic technologies has been shown to be unsatisfactory with percentage error of agreement with bolus thermodilution (% error) of 40% to 50%. Simultaneous measurement and averaging by two or more technologies may reduce random measurement scatter and improve precision. This concept, called the hybrid method, was tested by comparing accuracy and precision of measurement relative to bolus thermodilution using combinations of three component methods. Thirty patients scheduled for either elective cardiac surgery or liver transplantation were studied. Agreement with simultaneous bolus thermodilution of hybrid combinations of continuous thermodilution (Qt_CCO) or Vigeleo^TM/FloTrac^TM pulse contour measurement (Qt_FT) with pulmonary Capnotracking (Qt_CO2) was assessed pre- and post-cardiopulmonary bypass or pre- and post-reperfusion of the donor liver and compared with that of the component methods alone. Hybridisation of Qt_CO2 (% error 42.2) and Qt_CCO (% error 51.3) achieved significantly better precision (% error 31.3) than the component methods (P=0.0004) and (P=0.0195). Due to poor inherent precision of Qt_FT (% error 82.8), hybrid combination of Qt_FT with Qt_CO2 did not result in better precision than Qt_CO2 alone. Hybrid measurement can approach a 30% error, which is recommended as the upper limit for acceptability. This is a practical option where at least one component method, such as Capnotracking, is automated and does not increase the cost or complexity of the measurement process.

Statins are thought to potentially impair glucose metabolism, increasing plasma glucose concentration. The effect of prolonged statin use on glucose metabolism among outpatients is thus well established. However, the impact of statin use on glucose concentrations and insulin requirements during surgery remains poorly characterised and may very well differ considering the substantial hyperglycaemic stress response to surgery. We conducted a study to test the hypothesis that patients taking statins preoperatively require more intraoperative insulin than non-users. We analysed 173 adults having major non-cardiac surgery who participated in the Dexamethasone, Light Anaesthesia and Tight Glucose Control Trial between 2007 and 2010. We compared statin and non-statin users on total amount of intraoperative insulin to maintain plasma glucose concentration within 4.4 to 6.1 mmol/l using the inverse propensity score weighting method. Sixty-seven patients were statin users and 106 were non-statin users. The estimated ratio of geometric means between the statin users and the non-users was 1.45 (95% confidence interval: 0.93, 2.26, statin versus non-statin, P=0.11). The total amount of intraoperative insulin usage did not differ significantly among patients taking different types of statins (P=0.50). While the total amount of intraoperative insulin used was not statistically different between the statin users and non-users, we observed a potentially important trend toward insulin resistance intraoperatively among statin users during major non-cardiac surgery. This result is consistent with non-operative settings and cardiac surgery. Further investigation is essential to determine whether this effect is real and, if so, determine which specific statins are more associated with insulin resistance.

The aim of our study is to determine the short-term and long-term outcomes of intensive care unit (ICU) patients with minor troponin elevations. The retrospective study compared ICU patients with peak troponin elevation less than 0.1 ng/ml to those with only negative tests during their hospital stay. Data were gathered from ICUs at Beth Israel Deaconess Medical Center between 2001 and 2008. A total of 4224 patients (2547 controls and 1677 positives) were analysed. The primary outcome was mortality at one year. Secondary outcomes were 30-day mortality and hospital and ICU lengths of stay. After adjusting for age, sex, Simplified Acute Physiology Score, Sequential Organ Failure Assessment and combined Elixhauser score, we found that minor troponin elevations (peak troponin elevation between 0.01 and 0.09 ng/ml) were associated with a higher one-year mortality (Hazard Ratio 1.22, P <0.001 for binary troponin presence; Hazard Ratio 1.03, P <0.001 for each 0.01 ng/ml troponin increment). This relationship held for the subgroup of seven-day post-discharge survivors (Hazard Ratio 1.26, P <0.001). Minor elevations of troponin also significantly increased the net reclassification index over traditional risk markers for mortality prediction (net reclassification score 0.12, P <0.001). Minor troponin elevation was also associated with 30-day mortality (odds ratio 1.33, P=0.003). Importantly, troponin testing did not increase the adjusted mortality odds (P=0.9). Minor elevations in troponin substantially increase one-year, all-cause mortality in a stepwise fashion; it was also independently associated with 30-day mortality. We propose that minor elevations in troponin should not be regarded as clinically unimportant, but rather be included as a prognostic element if measured. We recommend prospective ICU studies to assess prognostic value of routine troponin determination.

There is some evidence that propofol may reduce acute postoperative pain; however, the results are inconsistent. Furthermore, there is a paucity of information about the type of anaesthesia and chronic pain. This study was designed to evaluate the hypothesis that propofol reduces acute and chronic postoperative pain compared with sevoflurane. In a randomised, prospective, double-blind trial, we assigned 80 patients having open total abdominal hysterectomy surgery to anaesthesia with either sevoflurane or propofol. Anaesthesia was titrated to clinical needs and bispectral index values to between 40 and 60. Postoperative pain was managed with pethidine and diclofenac. Acute postoperative pain for 24 hours and chronic postoperative pain at one and three months after surgery were evaluated. The Hospital Anxiety and Depression Scale was used to evaluate patient anxiety and depression after one and three months. There were no significant differences between the groups for opioid consumption or opioid-induced side-effects. Pain scores in the first four hours were significantly higher in the sevoflurane group. Persistent surgical pain was observed less frequently (7 out of 40 patients in the propofol group and 21 out of 40 in the sevoflurane group at three months post-surgery, P <0.01) and pain scores were lower at one and three months in the propofol group (0.78±0.55 versus 2.23±0.73 for the sevoflurane group at three months post-surgery, P <0.01). Anxiety and depression scores were significantly lower in the propofol group at three months. In this study, general anaesthesia with propofol was associated with reduced early acute postoperative and persistent pain, compared to sevoflurane-based anaesthesia, among patients undergoing open abdominal hysterectomy.

Simulation has been advocated as a useful training tool, and specific manikin simulators have been developed for use in this role. Debriefing and repetition have been identified as key to achieving educational goals and, while the technical features of manikin simulators can influence simulation outcomes, their cost and infrastructure requirements reduce their suitability for smaller healthcare facilities. We describe a local solution using biomedical calibration machines and modified basic manikins already available in our institution to form a high-fidelity anaesthetic simulator at minimal cost. This was effective in running high-fidelity, team-based in situ simulations and ‘can't intubate, can't oxygenate’ assessments for anaesthetic trainees. Though equipment in other centres may differ both in availability or suitability for simulation, the option we describe or similar may offer a low-cost solution for peripheral centres to run limited high-fidelity scenarios on a regular basis.

In this study, two commercially available quantitative neuromuscular function monitoring techniques, electromyography (EMG) and kinemyography (KMG), were compared with respect to repeatability and accuracy during late recovery from neuromuscular blockade. Train-of-four (TOF) ratios were recorded in 30 patients using KMG and EMG at the adductor pollicis muscle. Measurements were taken on the same hand using the Datex-Ohmeda NeuroMuscular Transmission monitor (GE Healthcare, Helsinki, Finland). Instrumental precision was evaluated using the coefficient of repeatability, while accuracy was assessed using the bias and limits of agreement. The coefficients of repeatability were similar for both techniques (0.035 for KMG and 0.043 for EMG), indicating a similar level of precision. KMG overestimated the TOF ratios measured with EMG with a bias of 0.11 (95% limits of agreement: −0.13 to 0.35). At a TOF ratio of 0.90 the bias was 0.08 (95% limits of agreement: −0.08 to 0.25). This means that at a TOF ratio of 0.90 measured with KMG will be approximately equivalent to a TOF ratio of 0.80 measured with EMG at the adductor pollicis muscle, but it may indeed be as low as 0.65 or as high as 1.00. Therefore, TOF ratios measured by KMG and EMG cannot be used interchangeably.

Platypnoea-orthodeoxia syndrome has the pathognomonic clinical findings of dyspnoea and arterial hypoxaemia relieved by recumbency. We report on a patient who presented with platypnoea-orthodeoxia syndrome post laparoscopic surgery. Platypnoea-orthodeoxia syndrome is an important diagnosis to consider when investigating hypoxia without an obvious cause.