Hypoglycaemia in hospital: a preventable killer?

People with known diabetes admitted to hospital have a longer length of stay and higher mortality rates [1]. In addition to this, many patients without known diabetes have elevated glucose levels on admission, which is a marker for poor outcomes in almost all scenarios [2–4]. Through vehicles such as the National Diabetes Inpatient Audit, there has been increased visibility and greater focus on identifying and treating people with hyperglycaemia in hospital. There is, however, an ongoing debate about optimum glucose targets in different populations and the best means of achieving glucose control. In most cases, dysglycaemia is managed with insulin, which is usually prescribed and adjusted by junior doctors, who report low confidence in managing diabetes [5]. Hypoglycaemia has long been recognized as the major side effect of insulin therapy, but its dangers are only now being realized. Analysis of large datasets shows that up to 26% of people admitted to hospital have hypoglycaemia, and this mainly occurs overnight [6]. Hypoglycaemia in hospital is also associated with longer length of stay and increased risk of mortality in the year after admission [7]. Whether or not this is a causal effect is unknown, as the risk of hypoglycaemia may merely be a marker of the underlying disease processes causing the mortality, rather than a direct contributor to mortality. In this issue of Diabetic Medicine, Rajendran and Rayman [8] survey adverse events related to inpatient hypoglycaemia. Twelve serious events (in 41 Trusts) were identified (defined as causing death, a cardiac or cerebral event or a fall resulting in permanent physical injury or fracture) and three of these events resulted in deaths. Insulin was implicated in 10 (83%) events and, in three events, hypoglycaemia followed the use of an i.v. infusion of insulin and dextrose to treat hyperkalaemia. Only one person had diabetes. It is important to recognize the limitations of this study. Responses were received from only 42 out of 142 trusts invited to participate. Although the absence of data from over three quarters of the hospitals surveyed suggests that there may well be a large number of unreported events out there, it is also possible that the study captured most of the severe events because of reporting bias. This study also highlights the dangers of insulin-dextrose infusions in patients with hyperkalaemia. These patients are usually ill and may also have acute kidney injury, which prolongs insulin action time and reduces counter-regulatory protection from hypoglycaemia. The variability in treatment response between patients makes careful glucose monitoring a necessity [9]. Insulin has been highlighted as a high alert medication [10]. A review of the National Patient Safety Agency incidents (November 2003 to November 2009) revealed a total of 16 600 incidents involving insulin. Of these incidents, 24% involved harm to patients and 18 incidents resulted in fatal and severe outcomes [11]. Errors occurred at all stages of the medication journey, including prescription, dispensing, administration and monitoring. These incidents are regarded as unacceptable, eminently avoidable and preventable. Causes include using the abbreviation ‘u’ instead of ‘units’, failure to administer insulin using insulin-specific devices, failure to give insulin when correctly prescribed, or issues involving unclear or misinterpreted verbal instruction about the use of insulin. Gupta and Cook [12] reported that the unintentional use of a glucose-containing solution has resulted in artifactually high glucose Cover image: Insulin crystals by light microscopy. Credit: Alfred Pasieka/Science Photo Library. Insulin is a common cause of hypoglycaemia in a hospital setting.