Characteristics of hypoglycaemic episodes in diabetic patients treated at the emergency room of a Portuguese tertiary centre, 2012‐2016

Abstract Objective To analyse changes in the characteristics of hypoglycaemic episodes treated in the emergency room of a tertiary hospital in Portugal between 2012 and 2016. Research Design and Methods We retrospectively analysed all emergency room reports for patients discharged with a diagnosis of hypoglycaemia between 2012 and 2016 and analysed demographic characteristics, type of diabetes and treatments, causes of hypoglycaemia and discharge destination. Patients without diabetes were excluded. Results In total, 676 hypoglycaemic episodes were analysed. Most patients were female (59%) and the median age of the patients was 71 years (interquartile range, 57‐81). The proportion of hypoglycaemic episodes relative to all emergency episodes decreased from 1.5% in 2012 to 1.0% in 2016 (P < .001). The proportion of patients with type 1 diabetes increased from 15.6% to 23.8%, while that of patients with type 2 diabetes decreased from 80.3% to 72.3% (nonsignificant differences). There was an increase in the use of insulin (67.1% to 85.4%, P = .02) and a decrease in the use of insulin secretagogues (26.6% to 11.5%, P = .03) over the study period. The rate of hospitalization dropped significantly from 11% in 2012 to 4.3% in 2015 and 5.4% in 2016 (P = .02). Conclusions Despite the increasing use of newer diabetes medications associated with a lower risk of hypoglycaemia, these episodes still require emergency care. The proportion of patients receiving insulin increased over the years, probably due to the slight increase in the prevalence of type 1 diabetes and the increasing replacement of secretagogues with insulin in type 2 diabetes.


| INTRODUC TI ON
Diabetes is a growing health concern. In 2017, an estimated 425 million people globally had diabetes and this figure is expected to reach 629 million by 2045. 1 Although many treatments exist for diabetes, just over 50% of patients with type 2 diabetes achieve the American Diabetes Association glycated haemoglobin (HbA1c) goal of <7.0% (53 mmol/mol). 2 Concern about hypoglycaemia is a barrier to optimal diabetes care, 3 and severe hypoglycaemia is the most serious adverse effect of insulin therapy in patients with type 1 diabetes. 4,5 Hypoglycaemia can cause acute harm to patients and others, especially when it causes car accidents, falls or other injuries. 6 Other severe consequences include vision loss, neurocognitive dysfunction, cerebrovascular disease and myocardial infarction. 7 Increased cardiovascular risk has also been reported, with an increased incidence of cardiac arrhythmias during acute hypoglycaemia 8 and an increased risk of thrombosis. 9 Hypoglycaemic episodes have also been linked to lower treatment satisfaction and adherence. 10 In the older population, severe hypoglycaemia is associated with a greater risk of dementia 11 and death. [12][13][14] Finally, hypoglycaemia poses a significant economic burden on both society and healthcare systems. 15 Hypoglycaemia is a common reason for emergency room (ER) visits and hospitalization. Between 1999 and 2011, over 400,000 people were admitted to US hospitals with hypoglycaemia, outnumbering those admitted for hyperglycaemia. 16 Based on data for 2013-2014, hypoglycaemic episodes accounted for 13.3% of all ER visits for adverse drug events in the United States. 17 In Portugal, the HIPOS-ER (Hypoglycemia in Portugal Observational Study-Emergency Room) study showed that 0.074% of type 2 diabetic patients required emergency care for severe hypoglycaemia. 18 The aim of this study was to describe the characteristics of hypoglycaemic episodes in diabetic patients discharged from a Portuguese ER with a diagnosis of hypoglycaemia between 2012 and 2016.

| RE S E ARCH DE S I G N AND ME THODS
We performed a retrospective observational study of all diabetesrelated episodes of hypoglycaemia treated at the ER of a tertiary hospital in Portugal between 2012 and 2016. We reviewed all ER reports for this period containing the following discharge ICD-9 (International Classification of Diseases, Ninth Revision) codes: 'specified hypoglycaemia', 'nonspecified hypoglycaemia' and 'hypoglycaemic coma'. Confirmed hypoglycaemic events only (capillary or venous plasma glucose level <70 mg/dL) were considered. ER admissions for nondiabetic hypoglycaemia were analysed separately from the episodes in people with diabetes. Admissions in which the report did not specify whether the patient had diabetes or not were excluded.
We recorded demographic characteristics, dependence on others, type of diabetes, treatments, levels of consciousness on admission, causes of hypoglycaemia, and destination at discharge from the ER for all patients included and compared data over the 5 years    are also more likely to have a higher burden of comorbidity and polypharmacy. Nevertheless, we did observe a decrease in the proportion of patients with type 2 diabetes presenting with hypoglycaemia requiring emergency care over the study period, probably because of the increasing use of antidiabetic drugs other than sulphonylureas.

| CON CLUS IONS
This hypothesis is supported by the steady decrease observed in ER patients with hypoglycaemia on sulphonylureas.
Conversely, we found an increase in the proportion of patients with type 1 diabetes admitted to the ER for hypoglycaemia. This is probably linked to the decrease observed for patients with type 2 diabetes and the higher prevalence of patients with type 1 diabetes under tight control with intensive insulin therapy.
The decrease in the proportion of patients taking sulfonylureas and the increase in those on insulin therapy have two likely explanations: the slight increase observed in hypoglycaemia among patients with type 1 diabetes and the increasing proportion of patients with type 2 diabetes being treated with insulin rather than insulin secretagogues. We also observed a decrease in the proportion of patients on NPH insulin and an increase in the use of insulin analogues, mainly as part of an intensive therapy regimen. It is well known that insulin analogues are associated with fewer hypoglycaemic events than traditional human insulins in both type 1 and type 2 diabetes. [21][22][23][24] Our results are therefore probably indicative of the increasing use of insulin analogues than of an increased risk of hypoglycaemia during this treatment.
In a survey of treatment-related severe hypoglycaemia in Japanese The overall hospitalization rate of 8.7% detected in our series for diabetic patients with a diagnosis of hypoglycaemia at discharge from the ER is much lower than the rates of 25% 26 and 29.3% 27 reported for ER visits in the United States and than the rate of hospitalizations reported on HIPOS-ER (44.1%). 18 Moreover, the annual hospitalization rates decreased steadily from 11% in 2012 to 4.3% in 2015, with a slight peak in 2016 (5.4%). We believe that this decrease can be explained by the lower use of sulphonylureas, especially long-acting, in the later years, as hypoglycaemic patients on these drugs were more likely to be hospitalized in both our series and the HIPOS-ER series. 18 As ours is a retrospective study, the conclusions that can be drawn are limited. One limitation is that the cases of hypoglycaemia were identified solely on the basis of discharge ICD-9 codes mentioned in the ER reports. We can speculate that these diagnoses are underestimated, as in our experience patients admitted for hypoglycaemia may be discharged with a diagnosis of a complication or an underlying disease that caused hypoglycaemia (eg trauma or aspiration pneumonia and acute kidney injury) rather than hypoglycaemia.
This can also explain our much lower rate of hospitalization comparing with HIPOS-ER. 18 In conclusion, our findings show that ER visits for hypoglycaemia are dropping, probably due to the growing use of new antidiabetic drugs to replace sulphonylureas in patients with type 2 diabetes.
Accordingly, more and more patients admitted to the ER for hypoglycaemia are on insulin therapy. We also detected an increasing use of insulin analogues compared with human insulins, but cannot confirm whether this change contributed somehow to the decrease in hypoglycaemic episodes seen at our ER. The advanced age of our patients did not vary significantly over the years studied, although we did observe a slight reduction in the proportion of frail patients and patients dependent on other people to administer their diabetes medication.

ACK N OWLED G EM ENTS
This paper had a nonrestrictive financial support of Novo Nordisk.

CO N FLI C T O F I NTE R E S T
The authors have no conflicts of interest to disclose.

AUTH O R S ' CO NTR I B UTI O N S
Catarina A. Pereira wrote the manuscript, researched and analysed data. Raquel Almeida researched data. Jorge Dores designed the study, contributed to the discussion and reviewed the manuscript.

E TH I C S A PPROVA L
All the preparation phases of the study were conducted in accordance with ethical and legal principles for data collection and statistical analysis. All clinical data were anonymized and analysed by an independent reviewer.

DATA AVA I L A B I L I T Y S TAT E M E N T
Author elects to not share data.