Peri‐operative management of diabetes: the need for a lead

At present, close to 5 million people in the UK (7.7% of the population) have a diagnosis of type-1 or type-2 diabetes mellitus and this number has doubled in the last 10 years; in addition, a further 20% of the UK population is at risk of type-2 diabetes 1, 2. Within the surgical population, it is estimated that 15% of patients have diabetes mellitus 3. Due to the rising prevalence of obesity and ageing of the population, diabetes will affect a growing proportion of the UK population in the future, and an increasing number of surgical patients will be affected by this metabolic state which is associated with significant adverse outcomes. Not only do patients with type-1 and type-2 diabetes see their life expectancy reduced by 20 years and 10 years, respectively 2, but they also have increased postoperative mortality primarily due to an increased incidence of complications, such as infection, acute myocardial infarction and acute renal failure 4. A large retrospective study involving > 44,000 patients scheduled for bariatric surgery showed that the presence of diabetes mellitus was associated with increased 30-day postoperative mortality (OR (95%CI) 2.58 (1.44–4.62) and 4.96 (2.74-8.97) for those treated with oral hypoglycaemic agents and insulin, respectively 5. More specifically, other authors have demonstrated in observational studies that peri-operative hyperglycaemia in non-cardiac surgery was associated with an increased incidence of complications and mortality 4, 6. Interestingly, when hyperglycaemia was recognised and treated with peri-operative insulin administration, complication and mortality rates were equivalent to those seen in non-diabetic patients with normal glycaemia 6. From an economic perspective, diabetes mellitus also affects the use of hospital resources, as diabetic patients often stay longer in hospital and intensive care 4, 6. Indeed, in an observational study of over 3000 non-cardiac surgery patients, Frisch et al. demonstrated that patients who were diabetic had a duration of stay 1.8 days longer than non-diabetic patients 4 (mean (SD) 8.8 (10.6) vs. 7 (10.8) days; p < 0.001). Therefore, in order to improve both patient- and economic-focused outcomes, it is necessary to aim for optimal glycaemic control throughout the peri-operative period in all patients with diabetes. Four years after publication of these recommendations, the National Confidential Enquiry into Patient Outcome and Death (NCEPOD) has recently published a report reviewing the quality of care provided to adult patients who had diabetes and underwent a surgical procedure 8. Of note, NCEPOD is an independent organisation, established more than 30 years ago, that uses a peer review and confidential enquiry method to examine the quality of care delivered in the peri-operative period 9. As per previous NCEPOD reviews this comprised an extensive retrospective case-note analysis, with 509 patient records included. Of note, there was an over-representation of emergency cases, and type-1 diabetic patients with a type-1:type-2 ratio of 1:3 (type 1-diabetic patients represent 10% of all those with diabetes 1). In addition, more than a quarter of the hospitals included were based in the independent, as opposed to the public, sector. These biases inherent to the methodology chosen by the authors do not undermine the validity of these results. Several key findings should be highlighted and are worthy of discussion. Less than half of referral letters had information on the management of the patient's diabetes and only 42% of these referral letters provided a HbA1c level within the previous 3 months despite national guidance recommending this. Part of this may be due to the high proportion of emergency referrals (43%) included in the analysis, where the patient may not have been referred by their usual primary care team. The need for the recognition and optimisation of chronic medical conditions by primary care, before the scheduling of surgery and pre-operative clinic attendance, is key in preventing unnecessary cancellations and delays in treatment provision 10. The importance of this has been highlighted in relation to other common comorbid conditions such as anaemia 11 and hypertension 12. With respect to hypertension, only around one-third of patients with diabetes had a blood pressure reading recorded in their referral letter. This report suggests that there is still scope for significant improvements to be made in the referral processes from primary care; however, this is a multifaceted, complex area which will take significant investment (both in terms of finance and personnel) in order to address. Initiatives such as the Royal College of Anaesthetists’ Peri-operative Medicine program will be one element in helping this process to evolve by providing a single point of contact for surgeons and general practitioners co-ordinating the care of complex patients 13. It is hoped that greater liaison with primary care, including provisional of educational programmes, will result in improved coordination between primary and secondary/tertiary care in management of long-term health conditions. Only 65% of patients had an HbA1c within 3 months of surgery and 16% of patients whose HbA1c was checked had a value outside the accepted range for elective surgery (> 69 mmol.mol−1). Again, the high number of non-elective cases may have contributed to these low compliance figures. However, the need for this target to be pre-requisite for elective surgery is worthy of consideration. This figure was first suggested in guidelines commissioned by NHS Diabetes 3 as one factor favouring good peri-operative diabetes control. More recent evidence has examined the association of HbA1c with postoperative mortality when controlling for peri-operative glucose level (the majority of earlier studies have examined hyperglycaemia and HbA1c in isolation) 14. This study showed that when controlling for blood glucose levels, HbA1c was not associated with increased 30-day mortality despite a strong relationship between HbA1c levels and peri-operative glucose levels. Thus, it may be that an elevated pre-operative HbA1c level should be viewed as a trigger for aggressive peri-operative glucose control, as opposed to deferral of surgery. This also applies to non-diabetic patients who are found to be hyperglycaemic, as high peri-operative glucose levels in this population are also associated with adverse outcomes 15. Aggressive peri-operative blood glucose control is also associated with a reduction in the risk of surgical-site infection 16; the data from this meta-analysis, however, did not take pre-operative HbA1c levels into account. Avoidance of postoperative nausea and vomiting (PONV) is of particular importance in patients with diabetes in order to get the patient eating and drinking again and, therefore, usual diabetic medication. The report judged that inadequate anti-emetic prophylaxis was given in 15% of cases, although the criteria for this judgement are not detailed. On review of the data provided it seems likely that multimodal anti-emetic prophylaxis was infrequently employed, as only 408 anti-emetic agents were used in 438 patients. This, in part, may be due to the omission of dexamethasone, which was only administered in 24% of cases. Dexamethasone increases glycaemic levels in both diabetic and non-diabetic patients by an average of 1.5 mmol.l−1, with a peak effect 4 h after administration 17. However, this does not appear to increase the risk of postoperative wound or systemic infection 18. Due to its prophylactic effect on PONV after general anaesthesia 19-21 or neuraxial anaesthesia with long-acting opioids 22, and its analgesic effect with either i.v. 23, 24 or perineural administration 24, 25, dexamethasone might be an interesting option. However, the decision as to whether to give it should be balanced between the anti-emetic and analgesic benefits and risk of hyperglycaemia. If dexamethasone is administered, then a low-dose should be chosen (i.e. 4 mg as opposed to ≥ 8 mg) as this is associated with lower postoperative blood glucose levels 26, and should be combined with another anti-emetic agent such as 5-HT3 antagonists, as a multi-modal approach is the key to success for effective anti-emetic prophylaxis. Despite over 90% of hospitals having protocols for the recognition and management of hypo- and hyperglycaemia, almost half the patients reviewed did not have their capillary blood glucose recorded intra-operatively and only 14% of patients had their blood glucose recorded in the recovery area. Significantly, one out of 20 diabetic patients in this report developed intra-operative hypoglycaemia 8. This complication, along with hyperglycaemia, is of great concern for the patient under general anaesthesia or the drowsy patient in the recovery room, due to lack of symptoms. The development and clinical implementation of continuous glucose monitoring devices continues to be disappointingly slow 27, leaving intermittent intra-operative sampling the only option at present. An alternative strategy to prevent occult intra-operative hypo- and hyperglycaemic episodes, would be to provide regional anaesthesia. Not only does the patient remain conscious and able to manifest any symptoms, but pain will be better controlled without opioid adverse-effects, such as PONV, that may worsen the condition. However, we should keep in mind that diabetes is associated with an increased risk of epidural abscess 28 or neuropathy 29, and prolonged block duration 30. The choice between general and regional anaesthesia should be individualised and based on a case-by-case situation, while properly weighing the risk-benefit ratio. Unfortunately, this report does not provide any information on the anaesthetic strategies and this, therefore, precludes further analysis. Based on these findings, the NCEPOD working group made similar recommendations to those in earlier guidelines (e.g. close glucose monitoring, safe and detailed handover of patients, avoidance of cancellation of surgery, prioritisation of patients with diabetes on the operating list) but it is always helpful to be reminded. More importantly, as only 28% of hospitals had a named clinical lead for peri-operative diabetes, the authors pointed out that a clinical lead in each hospital should be appointed to ensure that the pre-, intra- and postoperative management of diabetic patients will be optimal and performed following the state of the art and the latest recommendations. There was no recommendation as to what the base speciality of the clinical lead should be, so conceivably this could as be a physician, surgeon or anaesthetist. However, it may be that this responsibility falls within the remit of anaesthesia, due to its increasing role in pre-operative assessment and optimisation, where discussions with other medical specialities regarding the management of chronic disease conditions are initiated commonly. The need for co-ordinated multidisciplinary care has been a common theme in numerous previous NCEPOD reports, mentioned 15 times since 2001, but it is clear that it continues to be a challenge to implement. Part of the solution may be found in the ongoing work attempting to modernise the patient pathway to surgery 10. There appears to be little doubt that involvement of specialist diabetic teams early in these pathways will be of value in the peri-operative management of this complex disease, without forgetting that patient contribution and involvement in the management of this disease is of paramount importance 31. The NCEPOD group should be commended for leading this retrospective case-note survey, which enlightens the current practice in the UK 4 years after publication of guidelines. It is doubtless that the management of diabetic patients can be improved and that the skills and expertise to achieve this, such as diabetic nurse specialists and endocrinologists, are already in place but underutilised in the peri-operative setting. As such, the most important measure that each surgical department should take in the near future is to appoint a clinical lead to ensure that established recommendations are rigorously applied. Inadequate diabetic control needs to be recognised within the operating theatre environment as a risk factor for adverse outcomes that can be addressed with relatively simple clinical interventions. EA has received grants from the Swiss Academy for Anaesthesia Research (SACAR), Lausanne, Switzerland, from B. Braun Medical and from the Swiss National Science Foundation. EA has also received an honorarium from B. Braun Medical AG. MW is an editor of Anaesthesia.