Temporal changes in frequency of severe hypoglycemia treated by emergency medical services in types 1 and 2 diabetes: a population-based data-linkage cohort study.

Background: Almost 20 years ago, the frequencies of severe hypoglycemia requiring emergency medical treatment were reported in people with types 1 and 2 diabetes in the Tayside region of Scotland. With subsequent improvements in the treatment of diabetes, concurrent with changes in the provision of emergency medical care, a decline in the frequency of severe hypoglycemia could be anticipated. The present population-based data-linkage cohort study aimed to ascertain whether a temporal change has occurred in the incidence rates of hypoglycemia requiring emergency medical services in people with types 1 and 2 diabetes.
Methods: The study population comprised all people with diabetes in Tayside, Scotland over the period 1 January 2011 to 31 December 2012. Patients' data from different healthcare sources were linked anonymously to measure the incidence rates of hypoglycemia requiring emergency medical services that include treatment by ambulance staff and in hospital emergency departments, and necessitated hospital admission. These were compared with data recorded in 1997-1998 in the same region.
Results: In January 2011 to December 2012, 2029 people in Tayside had type 1 diabetes and 21,734 had type 2 diabetes, compared to 977 and 7678, respectively, in June 1997 to May 1998. In people with type 2 diabetes, the proportion treated with sulfonylureas had declined from 36.8 to 22.4% ( p  < 0.001), while insulin-treatment had increased from 11.7 to 18.7% ( p  < 0.001). The incidence rate of hypoglycemia requiring emergency medical treatment had significantly fallen from 0.115 (95% CI: 0.094-0.136) to 0.082 (0.073-0.092) events per person per year in type 1 diabetes ( p  < 0.001), and from 0.118 (0.095-0.141) to 0.037 (0.003-0.041) in insulin-treated type 2 diabetes ( p  = 0.008). However, the absolute annual number of hypoglycemia events requiring emergency treatment was 1.4-fold higher.
Conclusions: Although from 1998 to 2012 the incidences of hypoglycemia requiring emergency medical services appeared to have declined by a third in type 1 diabetes and by two thirds in insulin-treated type 2 diabetes, because the prevalence of diabetes was higher (2.7 fold), the number of severe hypoglycemia events requiring emergency medical treatment was greater.