BJA open最新文献

Background

Intravenous lipid emulsion is recognised as a therapy for rescue in cases of local anaesthetic toxicity, but its use in reversing overdose or toxicity related to other drugs remains the subject of debate. This in vitro study sought to expand our understanding of the importance of partitioning in determining the impact of intravenous lipid emulsion on aqueous free drug concentrations.

Methods

Twenty-seven drugs and associated metabolites were screened for the ability of intravenous lipid emulsion to reduce the amount of free drug in the aqueous phase, using specialised cassettes designed for this purpose. The relative amount of drug equilibrating across the membrane from plasma to phosphate-buffered saline was measured, using liquid chromatography–mass spectrometry, at a 6 h timepoint in plasma samples treated with intravenous lipid emulsion and paired, untreated controls.

Results

The data obtained were plotted against measures of partition (LogP and cLogD_7.4) and with log-transformed non-protein bound drug. There were significant inverse correlations between the capacity for intravenous lipid emulsion to reduce drug detected in the phosphate-buffered saline compartment and LogP and cLogD_7.4, and a direct association with log [non-protein-bound drug]. However, a number of drugs showed substantial variance between different plasma samples.

Conclusions

Modulation of free drug in the aqueous compartment is broadly predictable by the partition coefficient, although ramipril was identified to be an outlier in this regard. Further mechanistic and clinical exploration is merited to establish a standardised protocol for lipid emulsion therapy.

Background

The expansion of artificial intelligence (AI) within large language models (LLMs) has the potential to streamline healthcare delivery. Despite the increased use of LLMs, disparities in their performance particularly in different languages, remain underexplored. This study examines the quality of ChatGPT responses in English and Japanese, specifically to questions related to anaesthesiology.

Methods

Anaesthesiologists proficient in both languages were recruited as experts in this study. Ten frequently asked questions in anaesthesia were selected and translated for evaluation. Three non-sequential responses from ChatGPT were assessed for content quality (accuracy, comprehensiveness, and safety) and communication quality (understanding, empathy/tone, and ethics) by expert evaluators.

Results

Eight anaesthesiologists evaluated English and Japanese LLM responses. The overall quality for all questions combined was higher in English compared with Japanese responses. Content and communication quality were significantly higher in English compared with Japanese LLMs responses (both P<0.001) in all three responses. Comprehensiveness, safety, and understanding were higher scores in English LLM responses. In all three responses, more than half of the evaluators marked overall English responses as better than Japanese responses.

Conclusions

English LLM responses to anaesthesia-related frequently asked questions were superior in quality to Japanese responses when assessed by bilingual anaesthesia experts in this report. This study highlights the potential for language-related disparities in healthcare information and the need to improve the quality of AI responses in underrepresented languages. Future studies are needed to explore these disparities in other commonly spoken languages and to compare the performance of different LLMs.

Introduction

Severe traumatic brain injury affects ∼4500 per year across the UK. Most patients undergo a period of sedation to prevent secondary brain injury, however the optimal sedation target is unclear. This study aimed to assess the relationship between the electroencephalogram (EEG)-based Bispectral Index™ (BIS™) value and the clinical sedation score, along with other clinical outcomes.

Methods

Patients with severe traumatic brain injury in four UK ICUs were recruited to have blinded BIS data collected for a 24-h period while sedated on the ICU. Drug, physiological, and outcome data were recorded from the ICU record. Sedation management was at the discretion of the ICU clinical team.

Results

Twenty-six participants were recruited to the study. The mean BIS was 38 (inter-quartile range 29–44) and there was poor correlation between BIS and sedation score as a group (correlation coefficient 0.17, 95% confidence interval 0.08–0.26), however the spread in BIS values increased with decreasing sedation score. There was no statistically significant relationship between BIS and intracranial pressure, vasopressor use, osmotherapy use, or need for an additional sedative.

Conclusion

This study supports previous work showing that BIS decreases with decreasing sedation score. However, the variation in BIS values increased with deeper levels of clinical sedation. Patients may not be benefiting from the full potential of sedation in traumatic brain injury and further studies of sedation titrated to an EEG-based parameter are needed.

Clinical trial registration

NCT03575169.

Background

Group preoperative education is becoming standard care for patients preparing for surgery, alongside optimisation of exercise, diet, and wellbeing. Although patient education is essential, the effectiveness of group education programmes or ‘surgery schools’ as a means of delivery is unclear. This review examines whether attending group preoperative education improves patient outcomes.

Methods

We systematically reviewed studies of group perioperative education before major elective surgery. Observational or intervention studies with a baseline group or control arm were included. All outcomes reported were collected and, where possible, effect estimates were summarised using random effects meta-analysis.

Results

Twenty-seven studies reported on 48 different outcomes after group education. Overall, there was a 0.7 (95% confidence interval 0.27–1.13) day reduction in mean length of stay. The odds ratio for postoperative complications after abdominal surgery was 0.56 (95% confidence interval 0.36–0.85; nine studies). Patient-centred outcomes were grouped into themes. Most studies reported a benefit from group education, but only postoperative physical impairment, pain, knowledge, activation, preoperative anxiety, and some elements of quality of life were statistically significant.

Conclusion

This review presents a summary of published evidence available for group preoperative education. While these data lend support for such programmes, there is a need for adequately powered prospective studies to evaluate the effectiveness of preoperative education on clinical outcomes and to evaluate whether behaviour change is sustained. Furthermore, the content, timing and mode of delivery, and evaluation measures of preoperative education require standardisation.

Systematic review protocol

PROSPERO (166297).

Background

Patients are increasingly using artificial intelligence (AI) chatbots to seek answers to medical queries.

Methods

Ten frequently asked questions in anaesthesia were posed to three AI chatbots: ChatGPT4 (OpenAI), Bard (Google), and Bing Chat (Microsoft). Each chatbot's answers were evaluated in a randomised, blinded order by five residency programme directors from 15 medical institutions in the USA. Three medical content quality categories (accuracy, comprehensiveness, safety) and three communication quality categories (understandability, empathy/respect, and ethics) were scored between 1 and 5 (1 representing worst, 5 representing best).

Results

ChatGPT4 and Bard outperformed Bing Chat (median [inter-quartile range] scores: 4 [3–4], 4 [3–4], and 3 [2–4], respectively; P<0.001 with all metrics combined). All AI chatbots performed poorly in accuracy (score of ≥4 by 58%, 48%, and 36% of experts for ChatGPT4, Bard, and Bing Chat, respectively), comprehensiveness (score ≥4 by 42%, 30%, and 12% of experts for ChatGPT4, Bard, and Bing Chat, respectively), and safety (score ≥4 by 50%, 40%, and 28% of experts for ChatGPT4, Bard, and Bing Chat, respectively). Notably, answers from ChatGPT4, Bard, and Bing Chat differed statistically in comprehensiveness (ChatGPT4, 3 [2–4] vs Bing Chat, 2 [2–3], P<0.001; and Bard 3 [2–4] vs Bing Chat, 2 [2–3], P=0.002). All large language model chatbots performed well with no statistical difference for understandability (P=0.24), empathy (P=0.032), and ethics (P=0.465).

Conclusions

In answering anaesthesia patient frequently asked questions, the chatbots perform well on communication metrics but are suboptimal for medical content metrics. Overall, ChatGPT4 and Bard were comparable to each other, both outperforming Bing Chat.

Background

Accurate real-time prediction of intraoperative duration can contribute to improved perioperative outcomes. We implemented a data pipeline for extraction of real-time data from nascent anaesthesia records and silently deployed a predictive machine learning (ML) algorithm.

Methods

Clinical variables were retrieved from the electronic health record via a third-party clinical decision support platform and contemporaneously ingested into a previously developed ML model. The model was trained using 3 months data, and performance was subsequently evaluated over 10 months using continuous ranked probability score.

Results

The ML model made 6 173 435 predictions on 62 142 procedures. Mean continuous ranked probability score for the ML model was 27.19 (standard error 0.016) min compared with 51.66 (standard error 0.029) min for the bias-corrected scheduled duration. Linear regression did not demonstrate performance drift over the testing period.

Conclusions

We implemented and silently deployed a real-time ML algorithm for predicting surgery duration. Prospective evaluation showed that model performance was preserved over a 10-month testing period.

Background

Hypothermia is neuroprotective after neonatal hypoxic-ischaemic brain injury. However, systemic cooling to hypothermic temperatures is a stressor and may reduce neuroprotection in awake pigs. We compared two experiments of global hypoxic-ischaemic injury in newborn pigs, in which one group received propofol–remifentanil and the other remained awake during post-insult hypothermia treatment.

Methods

In both studies, newborn pigs were anaesthetised using halothane during a 45-min global hypoxic-ischaemic insult induced by reducing Fio₂ and graded hypotension until a low-voltage <7 μV electroencephalogram was achieved. On reoxygenation, the pigs were randomly allocated to receive 24 h of normothermia or hypothermia. In the first study (n=18) anaesthesia was discontinued and the pigs' tracheas were extubated. In the second study (n=14) anaesthesia was continued using propofol and remifentanil. Brain injury was assessed after 72 h by classical global histopathology, Purkinje cell count, and apoptotic cell counts in the hippocampus and cerebellum.

Results

Global injury was nearly 10-fold greater in the awake group compared with the anaesthetised group (P=0.021). Hypothermia was neuroprotective in the anaesthetised pigs but not the awake pigs. In the hippocampus, the density of cleaved caspase-3-positive cells was increased in awake compared with anaesthetised pigs in normothermia. In the cerebellum, Purkinje cell density was reduced in the awake pigs irrespective of treatment, and the number of cleaved caspase-3-positive Purkinje cells was greatly increased in hypothermic awake pigs. We detected no difference in cleaved caspase-3 in the granular cell layer or microglial reactivity across the groups.

Conclusions

Our study provides novel insights into the significance of anaesthesia/sedation during hypothermia for achieving optimal neuroprotection.

In the early days of anaesthesia, the fasting period for liquids was kept short. By the mid-20th century ‘nil by mouth after midnight’ had become routine as the principles of the management of ‘full stomach’ emergencies were extended to include elective healthy patients. Back then, no distinction was made between the withholding of liquids and solids. Towards the end of the last century, recommendations of professional anaesthesiology bodies began to reduce the fasting time of clear liquids to 2 h. This reduction in fasting time was based on the understanding that gastric emptying of clear liquids is rapid, exponential, and proportional to the current filling state of the stomach. Furthermore, there was no evidence of a link between drinking clear liquids and the risk of aspiration. Indeed, most instances of aspiration are caused by failure to identify aspiration risk factors and adjust the anaesthetic technique accordingly. In contrast, long periods of liquid withdrawal cause discomfort and may also lead to serious postoperative complications. Despite this, more than two decades after the introduction of the 2 h limit, patients still fast for a median of up to 12 h before anaesthesia, mainly because of organisational issues. Therefore, some hospitals have decided to allow patients to drink clear liquids within 2 h of induction of anaesthesia. Well-designed clinical trials should investigate whether these concepts are safe in patients scheduled for anaesthesia or procedural sedation, focusing on both aspiration risk and complications of prolonged fasting.

Background

Local anaesthetics are widely used for their analgesic and anaesthetic properties in the perioperative setting, including surgical procedures to excise malignant tumours. Simultaneously, chemotherapeutic agents remain a cornerstone of cancer treatment, targeting rapidly dividing cancer cells to inhibit tumour growth. The potential interactions between these two drug classes have drawn increasing attention and there are oncological surgical contexts where their combined use could be considered. This review examines existing evidence regarding the interactions between local anaesthetics and chemotherapeutic agents, including biological mechanisms and clinical implications.

Methods

A systematic search of electronic databases was performed as per Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines. Selection criteria were designed to capture in vitro, in vivo, and clinical studies assessing interactions between local anaesthetics and a wide variety of chemotherapeutic agents. Screening and data extraction were performed independently by two reviewers. The data were synthesised using a narrative approach because of the anticipated heterogeneity of included studies.

Results

Initial searches yielded 1225 relevant articles for screening, of which 43 met the inclusion criteria. The interactions between local anaesthetics and chemotherapeutic agents were diverse and multifaceted. In vitro studies frequently demonstrated altered cytotoxicity profiles when these agents were combined, with variations depending on the specific drug combination and cancer cell type. Mechanistically, some interactions were attributed to modifications in efflux pump activity, tumour suppressor gene expression, or alterations in cellular signalling pathways associated with tumour promotion. A large majority of in vitro studies report potentially beneficial effects of local anaesthetics in terms of enhancing the antineoplastic activity of chemotherapeutic agents. In animal models, the combined administration of local anaesthetics and chemotherapeutic agents showed largely beneficial effects on tumour growth, metastasis, and overall survival. Notably, no clinical study examining the possible interactions of local anaesthetics and chemotherapy on cancer outcomes has been reported.

Conclusions

Reported preclinical interactions between local anaesthetics and chemotherapeutic agents are complex and encompass a spectrum of effects which are largely, although not uniformly, additive or synergistic. The clinical implications of these interactions remain unclear because of the lack of prospective trials. Nonetheless, the modulation of chemotherapy effects by local anaesthetics warrants further clinical investigation in the context of cancer surgery where they could be used together.

Clinical trial regis