Factors and economic burden of non-severe hypoglycemia among insulin-treated type 2 diabetes patients: a cross-sectional study

Abstract Objective This cross-sectional survey was performed to assess the prevalence, factors, and economic burden of non-severe hypoglycemia among insulin-treated type 2 diabetes (T2D) patients in northern Thailand. Methods Between April 2021 and August 2022, 600 participants were evaluated via structured questionnaires containing sociodemographic and clinical characteristics, medications, and economic burden. Patients were divided into two groups (having and not having non-severe hypoglycemia). Variables with a p value <.05 in the univariate model were included in the multivariate model. Results The percentage of non-severe hypoglycemia was 50.3% (302/600). Of all participants, the average age was 61.4 ± 26.0 years, 55.7% were female, 53.5% used premix insulin, and the average duration of diabetes was 16.1 ± 10.0 years. Multivariate logistic regression analysis indicated that age (OR = .96; p <.001), duration of diabetes (OR = 1.04; p <.001), BMI (OR = .95; p = .002), thiazolidinedione (OR = 1.56; p = .012) and insulin regimens were associated with having non-severe hypoglycemia. Compared to basal insulin, basal bolus (OR = 6.93; p = .001), basal plus (OR = 3.58; p <.001), and premix insulin (OR = 1.83; p =.003) were associated with hypoglycemia. Greater numbers of sick leave were found in the hypoglycemia group (14 vs 4 patients, p = .029). Conclusions These findings help to individuate those patients who are at higher risk of non-severe hypoglycemia in insulin-treated T2D patients. Compared to the non-hypoglycemia group, patients with hypoglycemia were younger, had longer diabetes duration, lower BMI, received thiazolidinedione and insulin regimens such as premix, basal plus, or basal bolus insulins, and more productivity loss.


Introduction
The global prevalence of type 2 diabetes (T2D) in adults was 10.5% in 2021 1 , compared to 8.9% from the 5th Thai national health examination survey in 2014 2 .T2DM is characterized by slow progressive loss of beta cell functions, about one third with type 2 diabetes eventually require and benefit from insulin 3,4 .According to insulin access in tertiary care hospitals, insulin use among Thai diabetes was 25.3% 5 .Systematic review and meta-analysis of population-based studies found that hypoglycemia was more prevalent among T2D patients who take insulin; for mild/moderate episodes the prevalence was 50% and incidence was 23 events per person-year 6 .Mild-to-moderate hypoglycemic events were reported at 1.9 events/patient-year among Thai diabetes patients after starting or switching to a new insulin treatment 7 .Decreased body mass index, fasting plasma glucose change between visits, longer duration of insulin use, combination therapies with sulfonylureas, history of hypoglycemia and severe hypoglycemia, hypoglycemia awareness, and the presence of congestive heart failure were significant predictors of the frequency of non-severe hypoglycemia in insulin-treated T2D 8 .Non-severe hypoglycemia events are associated with adverse clinical outcomes, lower healthrelated quality-of-life, increased burden of disease, loss-ofwork productivity, and out-of-pocket costs [9][10][11] .Hypoglycemic symptoms in Thai diabetes patients demonstrated significantly higher impairment for EQ-VAS and EQ-5D indexes than those who did not experience hypoglycemic symptoms 12 and this is associated with increased worry 13 .In addition, hypoglycemic events incurred costly outpatient treatment of THB 3,102 or USD 103 and inpatient treatment of THB 74,532 or USD 2,475 14 .Due to a paucity of evidence related to non-severe hypoglycemia events among insulin users in Thailand, this study aimed to assess the prevalence of non-severe hypoglycemia and associated risk factors, and economic outcomes of non-severe hypoglycemia among insulin-treated patients with T2D who visited a tertiary care setting in northern Thailand.

Study design and participants
This cross-sectional study was carried out at the Maharaj Nakorn Chiang Mai University-Affiliated Hospital in Chiang Mai, a city in the North of Thailand, from April 2021 to August 2022.The diagnosis of T2D is determined by physician's discretion in conventional clinical practice using clinical parameters including, insidious onset, older age of onset, non-ketosis prone, and responsive to oral diabetes medications during initial management.Patients with T1D characteristics including ketosis prone, acute onset, non-obese, usually young onset, and positive for islet autoantibody (only available in some patients) were excluded from the study.
Patients with T2D visiting the outpatient department (OPD) who met the following criteria were recruited.The inclusion criteria were as follows: (1) aged �18 years; (2) received insulin with or without oral antidiabetic agents at least twice within 6 months; (3) able to provide informed consent and communicate with researchers; and (4) willing to participate in this study.Subjects who were terminally ill, had significant co-morbidities, such as cancer, chronic pain, or those diagnosed with mental illness were excluded.The sample size was estimated using the Krejcie & Morgan formula (n where n ¼ sample size, N ¼ population size, e ¼ acceptable sampling error, v 2 ¼ chi-square of degree of freedom 1, p ¼ proportion of population with hypoglycemic event) 15 .Based on a report of the Ministry of Public Health 16 , the number of insulin-treated patients with T2D in Thailand was approximately 650,000.The target sample size was determined assuming a worst-case scenario proportion of patients (50%) reporting at least one hypoglycemic event during the observation period, and sampling error of ±4%.All values were plugged in the Krejcie & Morgan formula.The estimated sample size in this study was 600 patients.

Ethics approval and informed consent
This study was approved by the ethics committee of the Faculty of Medicine, Chiang Mai University (MED-2564-08014).All procedures were carried out in accordance with the applicable guidelines and regulations.The study protocol was explained to all subjects.Each patient in the study read and signed a written informed consent form after agreeing to participate in this study.

Data collection
Patients with T2D who met the eligibility criteria of this study were included in the study.After patients were willing to participate in this study and had signed their written informed consent, face-to-face interviews were performed when patients were waiting at the OPD clinics.All interviewers were trained to standardize the data collection process.The questionnaires consisted of three parts (Supplementary material): (1) general patient information, (2) healthcare resource use due to hypoglycemia occurrence, and (3) clinical data, laboratory data and drug treatment.The first two parts of questionnaires were collected from the patients, while the third part was collected from the patients' medical profile.

Assessment of sociodemographic variables
The sociodemographic characteristics included age, sex, healthcare beneficiary which was composed of three health insurance schemes in Thailand, and self-payment.Other sociodemographic variables were educational level, job, marital status, and monthly household income.All sociodemographic variables were directly collected from the patients.

Assessment of clinical variables
The clinical variables considered in this study included duration of diabetes, body mass index (BMI), having chronic diabetic complications such as micro-vascular disease, macrovascular disease, and heart failure, having comorbidities such as hypertension, dyslipidemia, and laboratory results in the past 6 months.BMI was calculated by dividing weight (kg) by height (m) squared.The various minor comorbidities of the patients that might contribute to the risk of hypoglycemia were also recorded and grouped as appropriated for later analysis as a risk factor associated with non-severe hypoglycemia.

Assessment of medication use
Hypoglycemic drugs, insulin regimens, and other medications such as antihypertensive drugs and antilipidemic drugs were included in the study.Data of medication use were assessed at the time of data collection from the patients' medical profile.

Assessment of hypoglycemia in patients with T2D
In this study, we focused on non-severe hypoglycemia which was evaluated using the questionnaire based on the definition of hypoglycemia from the International Operations Hypoglycemia Assessment Tool (IO HAT) study 17 and the Standards of Medical Care in Diabetes 2020 18 .Non-severe hypoglycemia (any event managed by the patient alone/ symptoms of hypoglycemia, e.g.sweating, shaking, headache, with or without a blood glucose measurement, or a low blood glucose measurement (Level 1, defined as a measurable glucose concentration <70 mg/dL (3.9 mmol/L) but �54 mg/dl (3 mmol/L) or Level 2 hypoglycemia defined as a blood glucose concentration <54 mg/dL (3.0 mmol/L) with or without symptoms, that the individual managed without assistance from another person.

Assessment of economic burden
Economic burden was evaluated in terms of direct costs and productivity loss.Direct costs included costs of SMBG, and non-severe hypoglycemia treatment in which patients could access to care at the physician clinic, drug store, or self-care at home.The expenditures incurred from SMBG use, and non-severe hypoglycemia treatment were evaluated.In addition, days of sick leave due to non-severe hypoglycemia were self-reported by patients on a questionnaire based on their memories.

Statistical analysis
All samples were categorized into dichotomous groups according to the occurrence of hypoglycemia.Descriptive statistics of the sociodemographic, clinical and lab variables, and drug treatment were performed.Mann-Whiney U test for non-normally distributed variables or Fisher's exact test, or Pearson's chisquare test, whichever was appropriate, was performed to analyze group differences.Logistic regression analysis was used to explore the factors associated with hypoglycemia.In univariate logistic regression analysis, variables with a p value <. 05

Sociodemographic characteristics of patients with T2D
A total of 600 patients with T2D who received insulin at least twice in the past 6 months were enrolled in this study.Of those total subjects, about half (302 patients) had at least one event of non-severe hypoglycemia in the past 3 months.The sociodemographic characteristics of all subjects and dichotomous groups (hypoglycemia and non-hypoglycemia) are shown in Table 1.The mean age of overall patients was 61.4 ± 26.0 years old.Most patients were female (n ¼ 334; 55.7%), married (n ¼ 404; 67.5%), had health insurance under Civil Servant Medical Benefit Scheme (CSMBS) (n ¼ 299; 49.8%), had education at the elementary school (n ¼ 242; 40.3%), were unemployed (n ¼ 205; 34.2%), and had household income in the range of 608-1,824 USD per month (n ¼ 277; 46.2%).Only average age and gender were statistically significant differences between hypoglycemia and non-hypoglycemia groups.Compared to the non-hypoglycemia group, patients in the hypoglycemia group were slightly younger and had a greater percentage of females.Other characteristics did not differ significantly between both groups.

Clinical characteristics of patients with T2D
Table 2 represents the clinical characteristics of all patients with T2D and dichotomous groups.The average duration of diabetes and body mass index of patients were 16.1 ± 10.0 years and 27.0 ± 5.7 kg/m 2 , respectively.Approximately half of the patients had minor complications (n ¼ 322; 53.7%), in which chronic kidney disease accounted for 40.0%(n ¼ 120).One fourth of the patients reported major complications (n ¼ 150; 25%).Of those major complications, 120 patients (20.0%) had coronary artery disease.Less than 10% of all patients had heart failure.About 95% had comorbidities, like dyslipidemia or

Medical treatment of patients with T2D
In addition to insulin use, most patients received two classes of hypoglycemic drugs (n ¼ 191; 31.8%).The most frequently used hypoglycemic drug was metformin (n ¼ 373; 62.2%), followed by thiazolidinediones (n ¼ 234; 39.0%), dipeptidyl peptidase-4 (DPP-4) inhibitors (n ¼ 181; 30.2%), and sodium-glucose cotransporter-2 (SGLT-2) inhibitors (n ¼ 162; 27.0%).The group with hypoglycemia had a significantly higher number of thiazolidinediones users than the group with non-hypoglycemia (p ¼ .045).Four available insulin regimens were in use with different proportions for both groups.The group with hypoglycemia was likely to use more basal bolus insulin and basal plus insulin regimens than the group with non-hypoglycemia (p < .001).It was found that the total dose per day of insulin use in the group with hypoglycemia was likely to be higher than that in the group with non-hypoglycemia.Except for diuretic drugs, other medications were not significantly different between both groups.All the details are shown in Table 3.

Economic burden of patients with T2D
Although hypoglycemia occurrence was not severe and was managed without assistance, the group with hypoglycemia yielded greater economic burden in terms of costs of daily SMBG, hypoglycemia treatment, the number of sick leave patients, and days of sick leave, compared to the group with non-hypoglycemia (Table 4).

Factors associated with hypoglycemia in patients with T2D
Table 5 shows the results of univariate and multivariate logistic regression analysis.Variables with a p value <.05 in the

Discussion
This study has confirmed that non-severe hypoglycemia is considerably prevalent among insulin-treated patients with T2D.Of a total of 600 patients, 302 patients (50.3%) presented at least one event of non-severe hypoglycemia in the past 3 months.The results were in line with the findings of the meta-analysis of 532,542 people with T2D on oral therapies and insulin, which indicated 50% prevalence of mildmoderate episodes for those on insulin 6 .With the various sociodemographic factors, clinical characteristics, and medication treatment, this study revealed that age, sex, duration of diabetes, BMI, dyslipidemia, thiazolidinedione use, insulin regimens like premix insulin, basal plus insulin, and basal bolus insulin were associated with hypoglycemia in patients  with T2D.Targeting for more tight glycemic control in younger age individuals might explain the more hypoglycemic events in this study.In addition, young patients are more likely to comply with the recommendations for intensive glycemic control 20 .A gender effect on the counterregulatory hormonal response that occurs during a hypoglycemic event has been observed in women with diabetes 21 and, consistent with this, observations from six randomized clinical trials found a greater level of hypoglycemia in women during insulin treatment 22 .Longer duration of diabetes and low BMI may reflect insulin deficiency 23 .Thus, the use of insulin treatments may be more intensive and can cause more hypoglycemic events 24 .Statin treatment in general is associated with reduced incidence of hypoglycemia 25 , which may explain the lower hypoglycemia in patient with dyslipidemia in this study.Thiazolidinediones (TZDs), also known as insulin sensitizers, when added to the therapeutic regimen of patients with T2D who are receiving insulin may increase the risk of hypoglycemia 26 .Intensified insulin treatment has been found as a risk factor of insulin-treated T2D, especially in the elderly population 27,28 .Higher incidence of hypoglycemia, both mild and severe, have been reported in patients with T2D receiving complex regimens such as premix or prandial insulin, compared to those receiving a simpler basal insulin strategy 29 .We found that non-severe hypoglycemia led to greater productivity loss.Ten more patients were absent from work due to hypoglycemia compared with those with nonhypoglycemia (14 patients vs 4 patients).The average days of sick leave are double for the group with hypoglycemia compared to the group with non-hypoglycemia.In Thailand, there is a paucity of real-world evidence presenting the impact of hypoglycemia on productivity loss in terms of absenteeism and days of sick leave.Hypoglycemic episodes in insulintreated patients with T2DM causes anxiety 30 , worry 13 , and fear of future hypoglycemic events 31,32 .These effects may affect the patient's ability to carry out day-to-day tasks including housework, social activities, sporting activities, and sleep 33,34 .Across countries, the mean (SD) work-time lost among patients who reported work-time loss was 84.3 min for a daytime and 169.6 min for a nocturnal event 35 .The estimated productivity loss per non-hypoglycemic episode due to absenteeism calculated based on the proportion of respondents reporting missed work, multiplied by hourly income and hours missed, ranged from $15.26 to $35.58 (USD) in Germany, $46.30 to $83.59 (USD) in the UK, $26.43 to $55. 16 (USD) in the USA, and $48.33 to $93.47 (USD) in France 35 .
Several limitations with this study should be mentioned.First, we could not establish causal relationships between research variables and hypoglycemia due to the cross-sectional design.Second, although the samples in this study were from the largest tertiary care University-Affiliated Hospital in the North of Thailand, the generalizability of the findings to the whole country might be limited due to difference in clinical practice, patients' income, patients' behaviors, and so on.Third, accuracy of reporting hypoglycemic events was based on the memory of participants, which may have some recall bias and influence the findings.Therefore, further studies should broaden to more representative populations across Thailand and use more objective methods to reduce recall bias.Fourth, the under report of hypoglycemic events due to the possibility of hypoglycemia unawareness in some patients cannot be excluded.Lastly, the total dose of insulin use per day was collected as shown in Table 3.However, the total dose of daily insulin use as units per kilogram body weight could not be collected in this study due to self-adjustment of insulin dose, especially in those who self-monitor blood glucose.

Conclusions
Hypoglycemia is a risk for patients with T2D being treated with insulin.Our findings demonstrated that the prevalence of hypoglycemia was 50.3%.Compared to the non-severe hypoglycemia group, patients with hypoglycemia were younger, more female, had longer diabetes duration, lower BMI, were likely to receive thiazolidinedione treatment, received some types of insulin regimens such as premix, basal plus, or basal bolus insulins, and had more productivity loss.

Declaration of funding
The study was an investigator-initiated study, funded by a grant from Sanofi-Aventis (Thailand) Limited (PC-2019-12615).The sponsor played no role in designing the study, collecting, and analyzing data and composing the manuscript.

Declaration of financial/other relationships
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.
were included in multivariate logistic regression analysis (forward LR method).All statistical analyses were carried out in STATA software version 14.0 (StataCorp.2015.Stata Statistical Software.Release 14. College Station, TX: StataCorp LP).A two-sided p value <.05 was considered statistically significant.Costs were presented as Thai baht (THB) and converted into US dollars (USD) as 32.88 THB/USD 19 .

Table 1 .
Sociodemographic characteristics of people with type 2 diabetes.

Table 2 .
Clinical characteristics of patients with type 2 diabetes.

Table 3 .
Medication treatment of patients with type 2 diabetes.
CURRENT MEDICAL RESEARCH AND OPINION

Table 5 .
Univariable and multivariable type 2 diabetic patients with hypoglycemia.
CURRENT MEDICAL RESEARCH AND OPINION