Hypoglycaemia Among Insulin-Treated Patients with Diabetes Evaluation of the United Arab Emirates cohort of the International Operations-Hypoglycaemia Assessment Tool study

Objectives
This study aimed to evaluate the incidence of hypoglycaemia among insulin-treated patients with type 1 diabetes mellitus (T1DM) or type 2 diabetes mellitus (T2DM) from the United Arab Emirates (UAE) cohort of the non-interventional International Operations-Hypoglycaemia Assessment Tool study.


Methods
This cross-sectional observational study took place at 25 patient care centres in the UAE from October 2014 to May 2015. All adult patients with T1DM or T2DM who had been treated with insulin for >12 months were included. Self-assessment questionnaires and patient diaries were used to determine the incidence of documented hypoglycaemia both prospectively (four weeks after baseline) and retrospectively (six months and four weeks before baseline for severe and non-severe hypoglycaemic events, respectively).


Results
A total of 325 patients were enrolled in the study, of which 82 (25.2%) had T1DM and 243 (74.8%) had T2DM. Among patients with T1DM, 71.4% reported hypoglycaemic events retrospectively, with an incidence rate (IR) of 102.8 events per patient-year (PY), while 95% reported hypoglycaemic events prospectively, with an IR of 63.1 events per PY. Additionally, 56.3% of patients with T2DM reported hypoglycaemic events retrospectively, with an IR of 42.2 events per PY, while 91.9% reported hypoglycaemic events prospectively, with an IR of 33.3 events per PY.


Conclusion
The prevalence and incidence of hypoglycaemia were high among insulin-treated patients with T1DM and T2DM in the UAE. Individualised glycaemic goals, patient education and blood glucose monitoring may help to reduce the incidence of hypoglycaemia in this population.

H ypoglycaemia is a condition in which blood glucose (BG) levels drop below normal levels (3.9-7.1 mmol/L). 1 Approximately 90% of all patients treated with insulin experience hypoglycaemic episodes/events. 2 Complex regimes and glycaemic restrictions have been shown to increase the risk of hypoglycaemia, with severe hypoglycaemic events associated with increased mortality and morbidity. 3Hypoglycaemia also carries an economic burden, impacting patient productivity and disease management and utilising valuable healthcare resources. 4ccording to a recent meta-analysis, hypoglycaemia is prevalent among patients with type 2 diabetes mellitus (T2DM), thus necessitating an individualised approach to treatment and patient education. 5The Canadian Hypoglycaemia Assessment Tool Program revealed high rates of hypoglycaemia among insulin-treated patients; the researchers advocated for the identification of high-risk insulin-treated patients in order to reduce the incidence of hypoglycaemia in this patient group. 6In Germany, a patient-reported study revealed high rates of nonsevere hypoglycaemic events among those with type 1 diabetes mellitus (T1DM) or insulin-treated T2DM. 7nfortunately, it can be difficult to establish the actual rate of hypoglycaemic events as the majority of published data are based on the results of randomised controlled clinical trials (RCTs), which usually investigate the clinical efficacy and safety of certain drugs. 8,9Moreover, patients at higher risk of hypoglycaemia are often excluded from RCTs. 10 In 2017, the International Operations-Hypoglycaemia Assessment Tool (IO-HAT) study was conducted to determine the prevalence and incidence rates (IRs) of hypoglycaemic events among 7,289 insulin-treated patients with T1DM and T2DM in Bangladesh, Colombia, Egypt, Indonesia, the Philippines, Singapore, South Africa, Turkey and the United Arab Emirates (UAE). 11The IO-HAT study was itself based on data gathered as part of a larger investigation of 27,585 adult patients from 24 countries. 12The current study aimed to determine the incidence of hypoglycaemia among the UAE cohort of the IO-HAT study. 11In addition, the patients' knowledge of and attitudes towards hypoglycaemia were assessed, as well as the relationship between the incidence of hypoglycaemia and insulin regimens and glycaemic control, as determined by baseline glycated haemoglobin (HbA1c) measurements.

Methods
This cross-sectional non-interventional study was conducted at 25 patient care centres in the UAE between October 2014 and May 2015.All patients with T1DM or T2DM who were ≥18 years old at baseline and who had been treated with an insulin regimen for >12 months were included.Patients were enrolled in the study according to a consecutive sampling method during routine clinical consultation with their healthcare providers.As per the guidelines of the American Diabetes Association, severe hypoglycaemia was defined as an event of confirmed hypoglycaemia requiring the assistance of another individual to actively administer carbohydrate or glucagon interventions or take resuscitative action. 13Non-severe events were defined as symptomatic events not requiring biological confirmation and managed by the patient alone.Nocturnal hypoglycaemia was defined as an event occurring between midnight and 6 am. 13 Self-reported events of hypoglycaemia were documented by participants using a two-part self-assessment questionnaire and patient diaries.The primary endpoint of the study was the percentage of patients experiencing at least one hypoglycaemic event during a four-week prospective period; events were recorded in patient diaries completed over a four week period from baseline.The secondary endpoints included the IRs of hypoglycaemia prospectively (four weeks after baseline) and retrospectively (six months and four weeks before baseline for severe and non-severe events, respectively); these were determined according to a two-part selfassessment questionnaire, with the first part completed by the participants at baseline and the second part completed four weeks later [Figure 1].The two-part self-assessment questionnaire and patient diaries were translated into local languages and all acquired data were subsequently translated back into English for analytical purposes.
Other secondary endpoints included the patients' knowledge, awareness and fear of and attitudes towards hypoglycaemia and the relationship between the incidence of hypoglycaemia and insulin regimens and gly- Results from an international survey on 7289 patients from nine countries. 11aemic control.Awareness of hypoglycaemia was evaluated according to the patients' responses to the question 'Do you have symptoms when you have low sugar levels?' , in which the response 'usually' denoted impaired awareness and 'occasionally' or 'never' denoted severely impaired awareness.Fear of hypoglycaemia was self-assessed by patients on a scale of 0-10, where 0 denoted 'not afraid at all' and 10 was 'absolutely terrified' .Glycaemic control was determined by baseline HbA1c measurements, with percentages of >9%, 7-9% and <7% indicating poor, suboptimal and good glycaemic control, respectively.Comprehensive details of the design of the IO-HAT study and the procedures used for the assessment of hypoglycaemia have been previously reported by Emral et al. 11 The percentages of patients experiencing at least one hypoglycaemic event during the four-week prospective period were calculated along with 95% confidence intervals (CIs).Rates of hypoglycaemia were deemed equivalent in the retrospective and prospective periods according to two-sided statistical tests with the level of statistical significance set at P ≤0.050.The IRs of hypoglycaemia were calculated as the total number of events per patient-year (PY) divided by the total follow-up time in PYs along with 95% CIs.The difference in the reported incidence of hypoglycaemia was calculated using a negative binomial regression model, including a single binary covariate for two periods (four weeks before baseline and four weeks after baseline), specifying a logged exposure time as the offset term and using robust standard error to adjust for repeated measurements and the potential dependence between patients sharing the same site (site-level clustering).
This study was approved by the local ethics committee at each patient care centre, including the Al Qassimi Hospital Research & Ethics Committee, Sharjah (#154/2014-09-21), the Dubai Scientific Research Ethics Committee, Dubai (#DSREC-12/2014-03), the Institutional Review Board/Research Ethics Committee of Sheikh Khalifa Medical City, Abu Dhabi (#REC-0-5.02.2015RS-352) and the Gulf Medical University Ethics Committee, Ajman (#NovoNordisk/1/30032015). All study procedures were conducted in accordance with the guidelines of the International Society for Pharmacoepidemiology and the ethical standards of the revised Declaration of Helsinki. 14All patients provided informed consent prior to their participation in the study.
Among patients with T1DM, the IR of any hypoglycaemic event was higher in the retrospective period than the prospective period, although this difference was not significant (102.8events per PY, 95% CI: 94.8-11.3events per PY versus 63.1 events per PY, 95% CI: 56.0-70.7 events per PY; P = 0.093).The IR of nocturnal hypoglycaemia among patients with T1DM was significantly higher in the retrospective period compared to the prospective period (59.2 events per PY, 95% CI: 52.7-66.2events per PY versus 12.6 events per PY, 95% CI: 9.4-16.4events per PY; P <0.001).However,  the IR of severe hypoglycaemia in this group showed a non-significant increase in the prospective period compared to the retrospective period (9.4 events per PY, 95% CI: 6.8-12.6 events per PY versus 3.2 events per PY, 95% CI: 2.6-3.8events per PY; P = 0.682) [Figure 2A].
Similarly, the IR of any hypoglycaemic events among patients with T2DM was slightly higher in the retrospective period compared to the prospective period (42.2 events per PY, 95% CI: 39.2-45.2events per PY versus 33.3 events per PY, 95% CI: 30.4-36.3 events per PY).In contrast, the IR of severe hypoglycaemia was higher in the prospective period compared to the retrospective period (10.6 events per PY, 95% CI: 9.0-12.4events per PY versus 2.6 events per PY, 95% CI: 2.3-2.9 events per PY); nevertheless, both of these differences were non-significant (P = 0.253 and 0.142, respectively).As with the T1DM patients, the IR of nocturnal hypoglycaemia among patients with T2DM was significantly higher in the retrospective period compared to the prospective period (14.5 events per PY, 95% CI: 12.7-16.4events per PY versus 7.2 events per PY, 95% CI: 5.9-8.7 events per PY; P <0.001) [Figure 2B].
The IRs of any, nocturnal and severe hypoglycaemic events were calculated according to insulin regimens.For T1DM patients, the greatest IR of any hypoglycaemic events was observed with the use of a premixed insulin regimen in the prospective period (208.7 events per PY).The greatest IR of nocturnal hypoglycaemic events occurred among patients using a combination of short-and long-acting insulin regimens in the retrospective period (72.1 events per PY).For severe hypoglycaemic events, the greatest IR was observed among patients undertaking a long-acting insulin regimen in the prospective period (58.7 events PPY).The IRs for all types of hypoglycaemic events by insulin regimen for patients with T2DM in both the retrospective and prospective periods is presented in Figure 3.
No correlation was found between the percentage of patients experiencing hypoglycaemic events and glycaemic control in the prospective and retrospective periods.Among patients with T1DM, the majority (73.7%) of any hypoglycaemic events in the four-week retrospective period was experienced by patients with suboptimal glycaemic control; however, the proportion of patients experiencing any hypoglycaemic events was higher in those with poor glycaemic control compared to those with good glycaemic control (63.6% versus 61.5%).Among patients with T2DM, 64.8% of patients with poor glycaemic control, 54.5% with good glycaemic control and 53% with suboptimal glycaemic control experienced any hypoglycaemic events.
In terms of their perceptions of how hypoglycaemic events impacted the medical system, 5.2% versus 13.6% of T1DM patients and 0.5% versus 4.3% of T2DM patients reported that the event required hospital admission in the prospective period compared to the retrospective period.In addition, T1DM patients more frequently reported that the event resulted in additional telephone contacts retrospectively (16.7% versus 8.6%), while more T2DM patients reported this prospectively (20.1% versus 16.1%).Furthermore, more patients with T1DM undertook increased BG monitoring following the event compared to patients with T2DM in both the retrospective (54.9% versus 36.2%) and prospective (51.7% versus 31.7%)periods.In the retrospective period, fewer patients with T2DM compared to those with T1DM reported that hypoglycaemic events resulted in their absence (11.2% versus 27%), late arrival (9.1% versus 17.5%) or early departure (6.3% versus 15.9%) from work or studies [Table 2].Many patients with T1DM (84.2%) and T2DM (79.4%) had baseline knowledge of hypoglycaemia before being provided with the definition in the first part of the questionnaire.The number of patients defining hypoglycaemic events based on symptoms alone or in conjunction with BG measurements was comparable in both the T1DM and T2DM groups (47.6% versus 49% and 24.4% versus 29.2%, respectively).However, a small proportion of both T1DM and T2DM patients used only BG measurements to define hypoglycaemia (6.1% versus 2.9%).Fewer patients with T1DM had severely impaired awareness of hypoglycaemia compared to T2DM patients (1.2% versus 5.8%).The mean fear of hypoglycaemia score was comparable between the two groups (4.7 ± 3.6 versus 4.8 ± 3.7) [Table 3].

Discussion
According to the International Diabetes Federation, 425 million people had diabetes mellitus (DM) in 2017 and this number is expected to reach 629 million by 2045. 157][18][19] By 2045, it is predicted that 23.4% of the population between 20-79 years old in the UAE will have DM. 17,18Therefore, identifying the incidence of hypoglycaemic events is paramount in order to determine disease burden.However, recent studies have detected higher rates of hypoglycaemia in real-world settings when compared with the results of clinical trials. 5,8,20[11][12] As part of the larger IO-HAT study, the current study analysed self-reported rates of hypoglycaemic events among a cohort of insulin-treated DM patients from patient care centres in the UAE. 11Although higher rates of any or nocturnal hypoglycaemic events were reported in the current study during the retrospective period compared to the prospective period for both T1DM and T2DM patients, higher rates of severe hypoglycaemic events were reported prospectively in both periods; this may indicate that severe hypoglycaemic events were underreported by participants during the retrospective period.Gubitosi-Klug et al. similarly reported a higher incidence of severe hypoglycaemia in the prospective period among patients with T1DM. 21n the current study, the level of knowledge of the definition and symptoms of hypoglycaemia was similar to previously reported findings among diabetic individuals. 19,22This may be because such patients have individualised BG targets, greater experience with BG monitoring or have received patient education on this topic. 23Fear of hypoglycaemia can have major clinical repercussions for DM management.However, in the current study, low fear of hypoglycaemia was reported among both T1DM and T2DM patients, possibly due to adequate hypoglycaemia awareness and glycaemic control. 24Nevertheless, patients stated that hypoglycaemic events often negatively impacted their work or studies, with T1DM patients in particular more frequently reporting that such events resulted in absences, arriving late or leaving early compared to T1DM patients.Previous research also supports the notion that hypoglycaemia restricts social or functional activities, increases absenteeism, impairs punctuality and reduces productivity. 25,26revious studies have shown that all forms of hypoglycaemia are independent of levels of glycaemic control. 27,28Increased hypoglycaemic events in patients Excluding nine patients with T2DM for whom data for this category were unavailable.‡ Scored on a scale of 0-10, where 0 denoted 'not afraid at all' and 10 was 'absolutely terrified'.
with higher HbA1c levels may reflect poor self-care behaviours as well as aggressive efforts to improve their glycaemic control.Lash et al. advocate for increasing hypoglycaemia awareness, patient education and shared decision-making and the adoption of risk assessment and clinical decision-making support tools to support management of the condition. 29he current study is subject to certain limitations.During the enrolment of eligible patients, the possibility of volunteer bias could not be excluded.In addition, due to the retrospective design of the study, self-reported rates of hypoglycaemia may have been affected by recall bias; for example, patients may have under-or overreported severe hypoglycaemic events based on the definition alone, regardless of the severity of their symptoms or BG measurements. 3In addition, there may have been an overestimation in reported events while adjusting for recall bias, as reported rates of hypoglycaemic events were higher in patient diaries compared to their responses to the second part of the self-assessment questionnaire.Nevertheless, precautions were taken to ensure that rates of hypoglycaemia were not overreported by following the guidelines of the International Society for Pharmacoepidemiology. 14

Conclusion
To the best of the authors' knowledge, this is the first patient-reported dataset of hypoglycaemic events among the UAE cohort of the IO-HAT study.These findings highlight the high prevalence and incidence of hypoglycaemia among insulin-treated patients with DM and the potential impact of the condition on productivity, health and healthcare resource utilisation.Patient education and personalised glycaemic goals may help to lower the incidence of all forms of hypoglycaemia in this population, thus reducing associated costs to the healthcare system.

Figure 1 :
Figure 1: Diagram illustrating the design of the study.SH = severe hypoglycaemia; NSH = non-severe hypoglycaemia; SAQ = self-assessment questionnaire.Figure adapted with permission from: Emral R, Pathan F, Cortés CA, El-Hefnawy MH, Goh SY, Gómez AM, et al.Self-reported hypoglycemia in insulin-treated patients with diabetes:Results from an international survey on 7289 patients from nine countries.11

Figure 2 :
Figure 2: Incidence rates of hypoglycaemia among insulin-treated patients with (A) type 1 and (B) type 2 diabetes mellitus in the United Arab Emirates (N = 325).Hypoglycaemic events were assessed retrospectively over a four-week and six-month period (for nocturnal/ non-severe and severe hypoglycaemia, respectively) and prospectively over a four-week period.PY = patient-year; RR = risk ratio.

Figure 3 :
Figure 3: Incidence rates of (A) any, (B) nocturnal and (C) severe hypoglycaemia according to insulin regimen among patients with type 2 diabetes mellitus in the United Arab Emirates (N = 243).Hypoglycaemic events were assessed retrospectively over a four-week and sixmonth period (for nocturnal/non-severe and severe hypoglycaemia, respectively) and prospectively over a four-week period.PY = patient-year; SA = short-acting; LA = long-acting.
a c k n o w l e d g e m e n t sThe statistical analysis of the data was performed by Parexel International (Waltham, Massachusetts, USA).Medical writing and submission support was provided by Cognizant Technology Solutions Pvt. Ltd. (Pune, Maharashtra, India).Novo Nordisk A/S Pharmaceutical Company (Bagsvaerd, Denmark) was involved in the study conception and design, data collection, statistical analysis and interpretation and provided the materials, patients and resources.The decision to submit the article for publication was made by Novo Nordisk A/S Pharmaceutical Company.
c o n f l i c t o f i n t e r e s tProfessor Salah Abusnana and Dr Salem A. Beshyah received research grants and financial support for the submitted work from Novo Nordisk A/S Pharmaceutical Company.Professor Salah Abusnana, Dr Salem A. Beshyah, Dr Mahir Jallo, Dr Rima Tahhan and Dr Ravi Arora received payments for lectures and continuing medical education events as speakers.Dr Salem A. Beshyah, Dr Mahir Jallo and Dr Ravi Arora received reimbursements for travel and accommodation expenses for national and international conferences.Dr Hazem Aly and Dr Sagar Singhal are currently employed as medical advisors by the UAE branch of Novo Nordisk A/S Pharmaceutical Company.f u n d i n g This study was funded by Novo Nordisk A/S Pharmaceutical Company, Bagsvaerd, Denmark (HAT-IO #INS-4177).

Table 1 :
Characteristics of insulin-treated patients with type 1 and type 2 diabetes mellitus in the United Arab Emirates (N = 325) T1DM = type 1 diabetes mellitus; T2DM = type 2 diabetes mellitus; SD = standard deviation; HbA1c = glycated haemoglobin; FBG = fasting blood glucose; PPG = post-prandial glucose; BMI = body mass index; PVD = peripheral vascular disease; DPP = dipeptidyl peptidase; GLP = glucagon-like peptide; SGLT2 = sodium-glucose cotransporter-2; SA = short-acting; LA = long-acting.*Percentagesdonotadd up to 100% as some patients may have had more than one illness.†Percentagesarebased on the number of patients with available data.Hypoglycaemia Among Insulin-Treated Patients with Diabetes Evaluation of the United Arab Emirates cohort of the International Operations-Hypoglycaemia Assessment Tool study e450 | SQU Medical Journal, November 2018, Volume 18, Issue 4of insulin use was shorter among patients with T2DM compared to those with T1DM (6.8 ± 6.0 years versus 12.6 ± 8.6 years).However, mean baseline HbA1c percentages were similar in both groups (8.6 ± 1.6% versus 8.6 ± 1.9%) [Table1].A total of 262 patients completed the second part of the questionnaire after the prospective period had elapsed.Of these, 60 (22.9%) had T1DM and 202 (77.1%) had T2DM.

Table 2 :
Actions and perceptions following hypoglycaemic events among insulin-treated patients with type 1 and type 2 diabetes mellitus in the United Arab Emirates (N = 325) Hypoglycaemia Among Insulin-Treated Patients with Diabetes Evaluation of the United Arab Emirates cohort of the International Operations-Hypoglycaemia Assessment Tool study e452 | SQU Medical Journal, November 2018, Volume 18, Issue 4

Table 3 :
Knowledge, awareness and fear of hypoglycaemia among insulin-treated patients with type 1 and type 2 diabetes mellitus in the United Arab Emirates (N = 325) *Prior to the first self-assessment questionnaire being administered.†