By reading this article you should be able to: •Identify the key organisational components of 'opioid stewardship' programmes. •Identify key risk factors for persistent opioid use, opioid-induced ventilatory impairment and opioid diversion. •Develop effective written and verbal information for patients and caregivers on the safe use of opioids after discharge from hospital. •Describe the pharmacological properties of the ideal opioid to be prescribed at the time of discharge. Anna Simpson BA(Hons) PhD FRCA is a specialty registrar in anaesthesia in the Severn Deanery, UK. Nicholas Levy FRCA BSc is a consultant in anaesthesia and acute pain at West Suffolk Hospital. His major research interests are reducing harms from perioperative opioid use. He led the International multidisciplinary consensus statement on the prevention of opioid-related harm in the adult surgical patient. Edward R Mariano MD MAS FASA is professor and senior vice chair, Department of Anesthesiology, Perioperative and Pain Medicine at Stanford University School of Medicine, and chief of Anesthesiology and Perioperative Care, Veterans Affairs Palo Alto Health Care System. He has served as a director of the American Society of Regional Anesthesia and Pain Medicine and has chaired the American Society of Anesthesiologists' Pain Summit projects.

We are clinicians who are part of the ‘shielded’ or ‘extremely vulnerable’ group, as determined either by the government [1] or local policy. Whereas we appreciate the unprecedented and rapidly evolving nature of this situation, our group has been conspicuous by its absence from organisational guidance. The unexpected and frustratingly unique position we are in is strikingly different from our lifestyles of 2 months ago, andcurrently there is noexit strategy. Public Health England has offered no guidance with regard to employment in its shielding document [1, 2]. Whereas we have been making efforts to support frontline work from home, we have had mixed results. Barriers include physical distancing leading to reduced communication, lack of IT provision and confidentiality issues surrounding healthcare data. The guidance for those shieldings is simple: “Stay in your house, do not attend any gatherings, strictly avoid contact with those displaying symptoms” [1]. The initial shielding period is due to end on 30 June, though it is likely to be extended.With plans to re-introduce normalworkingpatterns across theNHS, we look to organisations, including the Royal College of Anaesthetists (RCoA), the Faculty of Intensive Care Medicine (FICM) and the Association of Anaesthetists, to help identify barriers and potential solutions to enable our safe, supported return to clinical duties. National guidance is limited. Pregnant staff have been offered an interpretation of the guidance by the RCoA/FICM/Royal College of Obstetricians and Gynaecologists, reflecting the associated risk involved with our clinical practice [3, 4], whereas those advised to shield have not as yet. The RCoA/FICM have acknowledged the uncertainty around how and when the shielding requirement may be lifted and that it will likely be on a case by case basis [5]. Whereas this acknowledges shielding, it lacks specific strategies/ advice for individuals, departments, Trusts and Deaneries going forward. A framework adaptable to individual circumstance that enables future planning should be developed. We strongly believe this is possible in order to support the inevitable upcoming transition period. Shielding anaesthetists and intensivists have some areas of concern; common themes include guilt, and a sense of being forgotten. Many departments have been incredibly supportive, but with little official guidance. There are feelings of anxiety and uncertainty regarding our futures, as anaesthetists and clinicians. Whether as a consultant, specialty doctor or trainee, working in their usual capacity has ceased for as long as the threat of COVID-19 remains. There will be training implications for many and some have concerns about returning to work, meaning discontinuation of clinical care or early retirement. Departments across the UK are now asking ‘how can we welcome shielding staff back to clinical settings safely’? We invite open discussion with key stakeholders on how we can facilitate our safe return towork. We are unsure how the guidance may evolve. Flexible return to work programmes and appropriate risk stratification are going to be key. Perhaps the option of working at clean sites could be explored [5]? We would welcome the chance to work with national bodies to address the following:

It is estimated by the International Diabetes Federation that 246 million adults worldwide have diabetes mellitus and the figure is expected to reach 380 million by 2025 [1]. Anaesthetists will be involved in the care of more diabetic patients as they present in increasing numbers for surgery as a result of the complications of diabetes. The cornerstone of metabolic control in the peri-operative period, except for Type II diabetics undergoing minor surgery, is the administration of intravenous (iv) glucose with potassium chloride and a variable insulin infusion. Standard anaesthetic and surgical texts recommend the use of 5% or 10% glucose at a rate of 125–83 ml.h−1 [2-6]. This corresponds to practice in nine out of 11 acute hospitals in the East Anglia region as shown by the authors’ recent audit (unpublished results). It is likely, therefore, that this regimen is common nationally. There are problems, however, with the uncritical acceptance of these recommendations, in particular the development of hyponatraemia. Excess free water is a well recognised complication of the use of glucose iv in non-diabetic patients [7] and has also been described with the use of sliding scale regimens in diabetic patients [6, 8, 9]. In non-diabetic patients the peri-operative use of glucose iv without sodium chloride would not be advocated, so the failure to emphasise the need for sodium in diabetic patients is surprising. This omission may have arisen, at least in part, as a consequence of the way iv regimens for diabetic patients developed over past decades. Peri-operative mortality amongst diabetic patients was high until the 1970s because of inadequate administration of insulin [10, 11]. A variety of methods to achieve glycaemic control were used in the 1970s with the insulin often administered subcutaneously. In 1979, in a seminal paper, Alberti and Thomas described a simple and safe method of achieving glycaemic control whereby glucose, insulin and potassium were infused at a fixed rate [12]. The ‘Alberti regimen’ rapidly became established practice at a time when infusion pumps were unreliable; the inclusion of glucose, insulin and potassium in the same bag of iv fluid ensured safety. A carbohydrate load of 180 g glucose per day was recommended to minimise catabolism (7.5 g.h−1). Alberti and Thomas also described the use of other iv fluids in conjunction with the glucose-insulin-potassium regimen, but lactate-containing solutions (such as Hartmann’s solution) were not recommended because they were thought to exacerbate the hyperglycaemia [12, 13]. The experimental data supporting the avoidance of lactate-containing solutions were weak [13] and there are strong theoretical reasons for suggesting that Hartmann’s solution is most unlikely to adversely affect glycaemic control. Since lactate is a 3-carbon compound and glucose is a 6-carbon compound, then in one litre of Hartmann’s solution the 29 mmol of lactate would yield 14.5 mmol of glucose at most. This assumes that the biosynthetic pathways of gluconeogenesis are 100% efficient, one litre is given instantaneously, and that no lactate is oxidised. The initial glucose distribution space is the extracellular fluid volume which in a 70 kg patient would be 12–15 litres. Thus, the maximum increase in glucose concentration with one litre of Hartmann’s would be about 1 mmol.l−1 and in clinical practice the effect on blood glucose will be much less. The longstanding irrational fear of using lactate-containing solutions has probably contributed significantly to hyponatraemia in diabetic patients. With the ‘Alberti regimen’ neither glucose nor insulin could be varied independently. If blood glucose was not adequately controlled, the bag of iv fluid had to be discarded and a new bag of glucose with the appropriate amount of insulin and potassium started. Although inherently safe, the practicalities of frequent changes of infusion made this regimen labour intensive which contributed to inadequate glycaemic control [14]. Variable rate insulin regimens became feasible in the 1980s as a result of reliable infusion pumps and the widespread availability of rapid accurate monitoring of blood glucose concentrations [14-16]. Insulin and glucose could be administered separately by infusion pumps with the rate of insulin infusion adjusted according to the circulating glucose value. This had become common practice in the Oxford region by 1993 when over 70% of anaesthetists used variable rate insulin infusions for glycaemic control during major surgery [15]. The use of separate glucose and insulin infusions can achieve excellent glycaemic control peri-operatively and provides substrate with insulin. However, the unthinking adoption of the regimen has inadvertently exposed diabetic patients to the potentially dangerous complication of hyponatraemia. Although this is unlikely to be more than a biochemical abnormality in most patients, symptoms of cerebral oedema may occur with lethargy and headache, seizures, coma and even death [17]. Grant and colleagues commented on the occurrence of hyponatraemia as a complication of variable rate insulin infusions and suggested the additional infusion of 0.9% sodium chloride in the event of hyponatraemia [6]. However, even though they recommended additional saline iv the rate of glucose infusion remained 100 ml.h−1. There is the obvious risk of fluid overload in these circumstances together with the practical difficulties of several iv infusions. Many studies have shown that hypotonic iv solutions predispose to hyponatraemia. In healthy volunteers, rapid infusion of 5% glucose iv resulted in significant hyponatraemia that had not returned to baseline values even after 6 h [18]. In sick hospital patients, the use of hypotonic iv fluids was found to be the major risk factor for the development of hyponatraemia [19]. A review of women who developed severe hyponatraemic encephalopathy after elective surgery concluded that the use of hypotonic fluids was the major cause [20]. There was a 27% mortality in this study and the survivors suffered severe permanent neurological impairment. Similar outcomes have been reported in further studies [21, 22]. The National Patient Safety Agency has recommended that hypotonic fluids should not be used in paediatric patients [23]. Indeed, a recent commentary on hyponatraemia advised strongly that hypotonic parenteral fluids should be avoided in all patients wherever possible [7]. We suggest that the ideal crystalloid solution for iv infusion in diabetic patients undergoing surgery should have the following properties: sufficient glucose to minimise catabolism and permit insulin infusion contain potassium and be compliant with safety recommendations [24] isotonicity not result in hyperchloraemic acidosis have widespread availability Fluids with some of these characteristics are available in other countries. For example, in USA 5% glucose with 0.45% sodium chloride and 20 mmol.l−1 potassium chloride has been advocated [25] and more complex balanced solutions with glucose are also available. Plasmalyte 148 and Dextrose contains 5% glucose, Na+ 140 mmol.l−1, K+ 5 mmol.l−1, Mg++ 3 mmol.l−1, Cl− 98 mmol.l−1, acetate 27 mmol.l−1 and gluconate 23 mmol.l−1 but is markedly hypertonic. Polionique B 66, a lactated Ringer’s solution with 1% glucose, is available in France but the glucose content is totally insufficient for diabetic patients [26]. The Raigmore Hospital in Inverness have chosen to run two iv fluids simultaneously, 10% glucose at 60 ml.h−1 and 0.9% sodium chloride with 20 mmol.l−1 potassium chloride at 60 ml.h−1 with an insulin infusion [27]. This practical solution decreases the likelihood of fluid overload but adds to the complexity of the iv regimens. At present, the best option for diabetic patients receiving an insulin infusion in the peri-operative period is 5% glucose in 0.45% sodium chloride solution with potassium 20 mmol.l−1. The lack of a readily available commercial product in the UK is a considerable handicap to managing these patients. Even this composition is less than ideal as it contains insufficient substrate if infused at 100–125 ml.h−1 and the amount of potassium may have to be varied. In the meantime we recommend that plasma sodium concentrations are measured frequently during the infusion of glucose and insulin so that hyponatraemia can be detected as early as possible and treated appropriately.

There are two established mechanisms, spike and modal inception, by which rotating stall is initiated in an axial flow compressor. Whilst the "Critical incidence hypothesis" and the "Zero slope criterion" are useful ideas in explaining the different stability boundaries for spikes and modes they do not provide the designer with a predictive tool. A detailed experimental investigation utilising a single-stage low-speed compressor is presented in which the aerodynamic environment of a rotor blade row is changed (rotor geometry is held fixed) so that it exhibited both spike and modal inception upon throttling into stall. The dominant mechanism of stall inception was found to be dependent on both the inlet flowfield and the downstream stator. The measurements are analysed and show that the meridional acceleration across the tip region of the rotor influences the mechanism by which rotating stall is incepted. This research is presented as a contribution towards the prediction of the stall inception mechanism.

AIMS/HYPOTHESIS:To investigate the pathways by which cyclic AMP (cAMP) stimulates glucagon-like peptide-1 (GLP-1) secretion, using the GLUTag enteroendocrine cell line. MATERIALS AND METHODS:GLP-1 release from GLUTag cells was measured in response to agents that increase cAMP, and single cells were studied by fluorescence calcium imaging and electrophysiology to evaluate the underlying pathways. RESULTS:Pituitary adenylate cyclase-activating polypeptide increased cAMP levels and stimulated GLP-1 release from GLUTag cells. Agents that increase cAMP levels, including forskolin plus 3-isobutyl-1-methylxanthine (fsk/IBMX), triggered a rise in the intracellular calcium concentration and enhanced the response to glucose by increasing both the number of cells responding to glucose and the magnitude of calcium responses in individual cells. Importantly, fsk/IBMX also stimulated GLP-1 release and intracellular calcium elevation even in the absence of nutrients. fsk/IBMX triggered membrane depolarisation and the firing of action potentials, associated with a +14 mV shift in the voltage-dependence of activation of hyperpolarisation-activated currents and the closure of a background potassium conductance. CONCLUSIONS/INTERPRETATION:We show here that cAMP elevation directly triggers GLP-1 release and enhances the secretory response to other stimuli like glucose, by modulating hyperpolarisation-activated currents and the background potassium current. cAMP-elevating pathways and the cAMP-modulated conductances in L cells present important targets for the development of therapeutic GLP-1 secretagogues.

In these experiments we have investigated the feasibility and accuracy of recording steady-state and dynamic changes in transmembrane potential noninvasively across an intact cell-attached patch using the current-clamp mode of a conventional patch-clamp amplifier. Using an equivalent circuit mimicking simultaneous whole-cell voltage-clamp and cell-attached current-clamp recordings we have defined both mathematically and experimentally the relationship between the membrane patch resistance, the seal resistance, and the fraction of the whole-cell potential recorded across an intact membrane patch. This analysis revealed a steep increase in the accuracy of recording of steady-state membrane potential as the seal/membrane ratio increases from 0. The recording accuracy approaches 100% as the seal/membrane ratio approaches infinity. Membrane potential measurements across intact cell-attached patches in rat basophilic leukemia cells and rat megakaryocytes revealed a surprisingly high degree of accuracy and demonstrated the ability of this noninvasive technique to follow dynamic changes in potential in nonexcitable cells.

Using a combination of fluorescence measurements of intracellular Ca2+ ion concentration ([Ca2+]i) and membrane potential we have investigated the sensitivity to serine/threonine phosphatase inhibition of Ca2+ entry stimulated by activation of the Ca2+ release-activated Ca2+ (CRAC) entry pathway in rat basophilic leukemia cells. In both suspension and adherent cells, addition of the type1/2A phosphatase inhibitor calyculin A, during activation of CRAC uptake, resulted in a fall in [Ca2+]i to near preactivation levels. Pre-treatment with calyculin A abolished the component of the Ca2+ rise associated with activation of CRAC uptake and inhibited Mn2+ entry, consistent with a requirement of phosphatase activity for activation of the pathway. Depletion of intracellular Ca2+ stores is accompanied by a large depolarisation which is absolutely dependent upon Ca2+ entry via the CRAC uptake pathway. Application of calyculin A or okadaic acid, a structurally unrelated phosphatase antagonist inhibits this depolarisation. Taken in concert, these data demonstrate a marked sensitivity of the CRAC entry pathway to inhibition by calyculin A and okadaic acid.

The incretin hormone, glucagon‐like peptide‐1 (GLP‐1) is released from intestinal L‐cells following food ingestion. Its secretion is triggered by a range of nutrients, including fats, carbohydrates and proteins. We reported previously that Na + ‐dependent glutamine uptake triggered electrical activity and GLP‐1 release from the L‐cell model line GLUTag. However, whereas alanine also triggered membrane depolarization and GLP‐1 secretion, the response was Na + independent. A range of alanine analogues, including d ‐alanine, β‐alanine, glycine and l ‐serine, but not d ‐serine, triggered similar depolarizing currents and elevation of intracellular [Ca 2 + ], a sensitivity profile suggesting the involvement of glycine receptors. In support of this idea, glycine‐induced currents and GLP‐1 release were blocked by strychnine, and currents showed a 58.5 mV shift in reversal potential per 10‐fold change in [Cl − ], consistent with the activation of a Cl − ‐selective current. GABA, an agonist of related Cl − channels, also triggered Cl − currents and secretion, which were sensitive to picrotoxin. GABA‐triggered [Ca 2 + ] i increments were abolished by bicuculline and partially impaired by (1,2,5,6‐tetrahydropyridine‐4‐yl)methylphosphinic acid (TPMPA), suggesting the involvement of both GABA A and GABA C receptors. Expression of GABA A , GABA C and glycine receptor subunits was confirmed by RT‐PCR. Glycine‐triggered GLP‐1 secretion was impaired by bumetanide but not bendrofluazide, suggesting that a high intracellular [Cl − ] maintained by Na + –K + –2Cl − cotransporters is necessary for the depolarizing response to glycine receptor ligands. Our results suggest that GABA and glycine stimulate electrical activity and GLP‐1 release from GLUTag cells by ligand‐gated ion channel activation, a mechanism that might be important in responses to endogenous ligands from the enteric nervous system or dietary sources.

Glucagon-like peptide 1 (GLP-1) secretion from intestinal L-cells is triggered by luminal nutrients. We reported previously that glucose-triggered GLP-1 release from the L-cell model GLUTag involves closure of ATP-sensitive K+ (K-ATP) channels. We show here that GLP-1 secretion and electrical activity of GLUTag cells is triggered not only by metabolizable sugars (glucose or fructose) but also by the nonmetabolizable monosaccharide methyl-alpha-glucopyranoside. Responses to glucose and methyl-alpha-glucopyranoside were impaired by the sodium-glucose cotransporter (SGLT) inhibitor phloridzin. SLGT1 and 3 were detected in GLUTag cells by RT-PCR. Whereas fructose closed K-ATP channels, methyl-alpha-glucopyranoside increased the membrane conductance and generated an inward current. Low concentrations of glucose and methyl-alpha-glucopyranoside also triggered small inward currents and enhanced the action potential frequency. We conclude that whereas low concentrations of metabolizable sugars trigger GLP-1 secretion via K-ATP channel closure, SGLT substrates generate small inward currents as a result of the electrogenic action of the transporter. This transporter-associated current can trigger electrical activity and secretion when the concentration of substrate is high or when outward currents are reduced by metabolic closure of the K-ATP channels. Electrogenic sugar entry via SGLTs provides a novel mechanism for glucose sensing by neuroendocrine cells.