METHODS
All patients with T1D who were seen at the Diabetes Centre of a regional hospital were invited to participate in this study. Exclusion criteria included patients who were pregnant, English illiterate or had significant cognitive impairment that negated taking an informed consent. Informed consent was obtained from subjects who agreed to participate in the study, with ethics approval obtained from the National Healthcare Group Domain Specific Review Board. This study consisted of 3 components: review of medical records, a detailed self-administered questionnaire (in English) and review of the blood glucose profile based on prospectively performed self-monitoring of blood glucose (SMBG) for 4 weeks.
Review of medical records included demographic data such as the patient’s age, gender and ethnicity, body mass index (BMI), age of diabetes diagnosis and the duration of diabetes, HbA1c values, results of insulin autoantibodies, as well as the insulin type and doses. The presence of diabetes-related complications – both microvascular complications such as retinopathy, nephropathy, neuropathy and macrovascular complications such as ischemic heart disease, stroke and peripheral vascular disease; were included. All prior hospitalizations or emergency department attendances for hypoglycaemia were also reviewed.
The questionnaire included several categories of important information. Firstly, diabetes self-management practices were assessed, which included frequency of SMBG (at least 3 times/day), methods for dosing of insulin (fixed or flexible insulin dosing), methods used to adjust insulin doses (based on carbohydrate intake and/or blood glucose), familiarity with carbohydrate counting and correction for hyperglycaemia. Secondly, prior diabetes education was assessed including the format, delivery and content covered. Thirdly, the Gold score and Clarke score, which are screening tools for IAH6,8,10, were assessed. The Gold score was obtained using a single question “do you know when your hypoglycaemic episodes are commencing?”, with the response graded on a 7-point Likert scale: 1 representing “always aware” and 7 representing “never aware”6. A Gold score of >= 4 indicates IAH6. The Clarke score comprises eight questions assessing the exposure to hypoglycaemia, symptomatic responses to hypoglycaemia and the glycaemic threshold8. A Clarke score of >= 4 indicates IAH8. Last but not least, the Problem Areas in Diabetes (PAID) questionnaire, a 20-item scale consisting of emotional problems commonly seen in people with diabetes mellitus, was administered.
The HYPO score, a useful indicator of the severity of problems with hypoglycaemia, was computed11. This score comprised of two components – the first component was based on historical recall of hypoglycaemic events requiring assistance during the preceding one year – assistance required to recognize hypoglycaemia, assistance required to treat hypoglycaemia, glucagon administration required for hypoglycaemia, and emergency medical services required for hypoglycaemia, in the last one year11. Severe hypoglycaemia was defined in our study as the occurrence of at least one of the above four events in the preceding one year. This is consistent with the American Diabetes Association definition of severe hypoglycaemia as severe cognitive impairment requiring external assistance for recovery2. The accuracy of self-recall of severe hypoglycaemia was confirmed by Pedersen et al, who demonstrated that 90% of subjects with confirmed episodes of severe hypoglycaemia during a prospective one year period of monitoring were able to successfully recall these episodes12. The second component of the HYPO score was calculated based on the prospective period of SMBG, as discussed in the next section.
Following administration of the questionnaire, subjects who consented were provided with a standardized glucometer and 150 test strips and instructed to perform SMBG before every meal and at bedtime, daily for 4 weeks. Clinically significant hypoglycaemia was defined as an episode in which capillary blood glucose was < 3 mmol/L during this prospective 4-week period of SMBG. A glucose level of < 3 mmol/L was chosen as it is sufficiently low to indicate serious, clinically important hypoglycaemia13. During all episodes in which blood glucose was < 3 mmol/L, subjects were asked to record the absence or presence of symptoms and whether these were autonomic or neuroglycopenic symptoms – this was the second component required for calculation of the HYPO score.
Statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS) software. Descriptive statistics were presented as median with interquartile range for quantitative variables and number (percentages) for qualitative variables. Univariate analysis was performed by using independent sample t-test, Pearson’s chi-square test, Mann-Whitney-U test and Fisher exact test whenever appropriate to compare demographic and clinical factors between subjects with impaired and intact awareness of hypoglycaemia. A p-value of < 0.05 was considered statistically significant and all p-values reported were two-sided.