SURGERY IN DIABETES
Target glucose – during surgery, blood glucose should be 7–11 mmol/litre. At normal levels, patients are too close to hypoglycaemia. At levels above 11 mmol/litre, urine output increases and dehydration may ensue. Fluids – any other fluids given during the surgical period should not contain glucose. Use of Hartmann’s solution (Ringer lactate) in patients with diabetes remains controversial. The lactate contained in this crystalloid is used for gluconeogenesis, particularly in starved or catabolic patients. In patients with type 2 diabetes, an infusion of Hartmann’s solution may cause blood glucose levels to rise significantly. If fluids have to be restricted, glucose may be given as a 20% or 50% solution. This must be administered via a central venous catheter, to avoid venous thrombosis. Electrolytes – potassium levels should be monitored regularly perioperatively. Serum potassium varies according to: • the effects of insulin, which promotes potassium uptake by muscle, liver and adipose tissue • dehydration, which may cause a shift in potassium from the intracellular to the extracellular space • acidosis, which leads to hydrogen and potassium exchange in the kidneys, potassium retention and hyperkalaemia. Most patients with normal renal function require 20 mmol potassium/litre fluid given, but the requirement is often higher in patients with diabetes. Analgesia – in the past, there was a view among anaesthetists that regional blockade (including spinal and epidural) was undesirable in diabetes. This originated from fear of aggravating (possibly latent) neuropathy, of uncontrolled hypotension in those with significant autonomic neuropathy, or of causing infection at the site of the block. This view is no longer valid. The advantages of regional blockade, which provides excellent analgesia and blunting of the stress response, outweigh any disadvantages in most patients with diabetes.
Surgery in Patients with Diabetes Simon R Heller Sean F Dinneen
Diabetes presents several special problems during surgery. Fasting causes particular problems in type 1 diabetes. Such patients need basal insulin to prevent ketosis, and develop hypoglycaemia without additional carbohydrate intake. Fasting is of little significance in type 2 diabetes, unless the patient has received oral hypoglycaemic agents. Metabolic changes include the following. • Increases in circulating adrenaline, adrenocorticotrophic hormone, cortisol and growth hormone aggravate insulin deficiency and insulin resistance. These changes are a normal response to surgery and proportional to the severity of the operation. They antagonize the actions of insulin and its secretion, resulting in catabolism with increased glycogenolysis, gluconeogenesis, proteolysis and lipolysis. In diabetes, the effects are compounded by insulin deficiency. • Gluconeogenesis from precursors such as lactate, pyruvate, alanine and glutamine is increased in the liver and kidney, and muscle and adipose tissue take up less glucose. The resultant hyperglycaemia is more pronounced in patients with diabetes than in non-diabetic patients. • Without insulin, lipolysis is stimulated and leads to ketogenesis. Plasma levels of free fatty acids, glycerol and ketone bodies increase, and metabolic acidosis may develop even in the presence of near-normal plasma glucose. All these changes are aggravated by some types of anaesthesia, particularly high doses of opiates or regional blockade. They increase insulin requirements in patients with type 1 diabetes, and may cause those with type 2 diabetes to become temporarily insulin-requiring. Recognizing hypoglycaemia may be difficult in unconscious patients. Subcutaneous insulin absorption is poor or unpredictable when peripheral vessels are constricted.
Emergency surgery Up to 5% of diabetic patients require emergency surgery at some time, usually because of infection. Danger results from the loss of control of diabetes and the onset of ketoacidosis provoked by the surgical illness. There may be confusion if diabetic ketoacidosis mimics an acute abdomen. Glycaemic control and acid–base balance should be thoroughly evaluated by blood glucose, electrolytes, blood gases and urinary ketones. Intravenous fluids should be given, with potassium and insulin as indicated. In up to 60% of patients with diabetic ketoacidosis, abdominal symptoms resolve following control of the diabetes and fluid therapy, avoiding the need for laparotomy. However, there is a further pitfall for the unwary – autonomic neuropathy may mask the symptoms and signs in a patient with diabetes with a true acute abdomen needing urgent laparotomy.
Principles of management The fundamental principle of surgical management in diabetes is that capillary blood glucose is measured regularly and accurately, and that these results are recorded and acted on. Most problems occur because staff have forgotten to measure blood glucose, or because very low values have been ignored or wrongly attributed to faulty meters.
Management strategies
Simon R Heller is Reader in Medicine at the University of Sheffield and Honorary Consultant Physician at the Northern General Hospital, Sheffield, UK. Sean F Dinneen is Consultant Diabetologist at Addenbrooke’s Hospital, Cambridge, UK.
MEDICINE
There are three strategies (Figure 1) for managing patients with diabetes undergoing surgery – omission of antidiabetic medication and careful monitoring, separate glucose and insulin infusions, and glucose/potassium/insulin (GKI) infusion.
66
© 2002 The Medicine Publishing Company Ltd
SURGERY IN DIABETES
Regimens for surgery in diabetes Type of diabetes • Type 1 (always treated with insulin, ketosis-prone) • Type 2 (diet/tablets or insulin, not prone to ketosis)
Minor Fast-and-check
Type of surgery Moderate Major Glucose–insulin or GKI Glucose–insulin or GKI
Emergency Glucose–insulin
Fast-and-check
Fast-and-check
Glucose–insulin
Glucose–insulin or GKI
1
Infusion rate adjustments in the glucose–insulin regimen Blood glucose (mmol/litre) • <4 • 4–6.9 • 7–10.9 • 11–16.9 • ≥ 17
Pump rate (ml/hour) Stop 1 2 3 4
Comments Check glucose infusion is running, re-check in 30 minutes; if still low, call doctor Two successive readings require a reduction in insulin infusion rate by 1 ml/hour Two successive readings require an increase in insulin infusion rate by 1 ml/hour Re-check in 1 hour; if still high, check insulin is running, call doctor
Check plasma potassium before commencing regimen, immediately postoperatively and at least daily during infusion. These rates may need adjusting in different situations, depending on the age, weight and condition of the patient; generally, the more ill the patient, the more insulin needed.
2
Omission of antidiabetic medication and regular monitoring (‘fast-and-check’ strategy) is suitable for: • short procedures (e.g. gastroscopy, dental extractions) in all patients • minor procedures (e.g. arthroscopy, dilatation and curettage) in all patients • moderate procedures (e.g. hernia repair, hysterectomy) in patients not prone to ketosis (i.e. those with type 2 diabetes). The patient is placed early on the operating list. The usual insulin injection or tablets are omitted, and capillary blood glucose is measured hourly using a visually-read strip or glucose meter. A single dose of soluble insulin (Actrapid), 6 units s.c., is given if levels rise to 17 mmol/litre or above. If blood glucose continues to rise in the following 2 hours, an intravenous regimen should be considered. Patients revert to the normal regimen when they are able to eat and drink. Separate glucose/insulin infusions are suitable for: • all emergency surgery • moderate or major surgery in type 1 diabetes • major surgery in type 2 diabetes. Patients should not be first on the list – they should be placed towards the end or in the afternoon, to allow a few hours for the regimen to shift blood glucose levels into the desirable range (7–11 mmol/litre). The glucose infusion comprises 10% dextrose, 500 ml, with potassium chloride, 10 mmol, given at a rate of 100 ml/hour. The infusion rate must be controlled using a pump (e.g. IVAC); it is hazardous to attempt to control the rate by eye. The insulin infusion comprises soluble insulin (Actrapid), 50 units, in 0.9% saline, 50 ml, delivered in a syringe driver. Capillary blood must be measured accurately every hour and the insulin pump rate adjusted as shown in Figure 2. GKI infusion is suitable for: • moderate or major surgery in type 1 diabetes • all major surgery in type 2 diabetes.
MEDICINE
An infusion of 10% dextrose, 500 ml, with potassium chloride, 10 mmol, and Actrapid insulin, 10 units, is given at a rate of 100 ml/hour, using a rate controller. Capillary blood glucose is measured precisely every hour. If blood glucose falls to 4–7 mmol/litre in two successive hours, less insulin is needed; the solution should be changed to 5 units/500 ml. If blood glucose rises to 17 mmol/litre or more in two successive hours, more insulin is needed; the solution should be changed to 20 units of insulin/500 ml. Plasma potassium must be checked before starting the regimen, immediately postoperatively, and at least daily during infusion. These recommendations may need adjusting depending on the age, weight and condition of the patient; generally, the more ill the patient, the more insulin needed. The suggested dextrose infusion rates are those required to control blood glucose. During prolonged fasts following major surgery, saline and other fluids are required depending on the type of surgery and the condition of the patient; these must be agreed between the anaesthetist and the surgical team.
Postoperative management Once the patient is eating and drinking again, he can return to his usual treatment. Patients with type 1 diabetes must take the first dose of subcutaneous insulin 60 minutes before any intravenous insulin is discontinued. This is to prevent them becoming ketotic during any interval between cessation of the infusion and the injection.
FURTHER READING Eldridge A J, Sear J W. Perioperative Management of Diabetic Patients. Anaesthesia 1996; 51: 45–51. Milaskiewicz R M, Hall G M. Diabetes and Anaesthesia: The Past Decade. Br J Anaesth 1992; 68: 198–206.
67
© 2002 The Medicine Publishing Company Ltd