The prevalence of diabetes mellitus (DM) worldwide is steadily increasing, and thus the treatment of DM has become an important social and economic issue. The Korean Diabetes Association (KDA) has been publishing the Diabetes Fact Sheet since 2012; according to the report published in May 2022, approximately six million Korean adults aged 19 years or older are estimated to have DM. Among patients with DM aged 30 years or older, 7.5% patients are estimated to use insulin [1]. Insulin treatment is necessary for all people with type 1 diabetes mellitus (T1DM) and a subgroup of people with type 2 diabetes mellitus (T2DM) who are no longer able to reach an appropriate glycemic target with oral antidiabetic medication alone or those who have hyperglycemic symptoms or catabolic features [2]. Patients with T1DM typically require initiation with multiple daily injections (basal [long acting] insulin and bolus [short acting] insulin) or continuous subcutaneous insulin infusion. The preferred regimen of insulin initiation in T2DM patients is starting basal insulin in addition to oral antidiabetic agents. If glycemic target is not achieved, bolus insulin can also be added to control postprandial glucose levels in T2DM patients.

After the first successful extraction of insulin from dog pancreas in 1921, the subsequent development of human recombinant insulin such as humulin R (rapid acting) and N (intermediate acting) was achieved [3]. The Diabetes Control and Complications Trial (DCCT) in 1993 found that the intensive treatment of diabetes delays the onset and slows the progression of microvascular complications in T1DM [4]. The development of short and long acting insulin analogs and their combination therapy has enabled more physiologic and intensive glycemic control without hypoglycemia, which is a major limiting factor of intensive glycemic control. Treatment with the analog insulins is associated with not only less hypoglycemia and weight gain but also lower glycosylated hemoglobin (HbA1c) compared with human insulin in T1DM patients [5]. The American Diabetes Association recently recommended a glucagon-like peptide-1 receptor agonist (GLP-1 RA), which is preferred to insulin when possible in T2DM, and there are diverse anti-diabetes treatment options [6]. Treatment intensification for patients who do not meet treatment goals should not be delayed and insulin is one of the important antidiabetic treatment options. Therefore, insulin treatment continues to play a necessary and important role for effective patient treatment. From this point of view, it is meaningful to investigate the current status of insulin use and the use of antidiabetic agents among insulin users.

In this study, we investigated the trends of insulin therapy in people with T1DM and T2DM in Korea. The trends of antidiabetic drug use among patients with T2DM who use insulin were also evaluated.

The number of insulin users with T2DM has continued to increase since 2002. Among insulin users, the proportion of patients who use basal analogs has increased rapidly, reaching 75.1% of total insulin users with T2DM in 2019. In addition, the proportion of patients who use multiple daily injections, which is considered an intensified insulin regimen, has increased steadily. Among insulin users with T2DM, the use of antidiabetic medications other than insulin continues to increase, and the proportion of patients receiving dual therapy in addition to insulin is especially increasing. The most common dual combination therapy is metformin plus DPP-4 inhibitors, reaching 26.1% of total insulin users in 2019, followed by metformin plus SGLT2 inhibitor and metformin plus sulfonylurea. Considering that the use of metformin plus sulfonylurea accounted for up to 19.2% of patients in 2014, the use of this combination is decreasing, while the use of DPP-4 inhibitor and SGLT2 inhibitor is increasing. In addition, the use of metformin plus GLP-1 RA is increasing. This trend is because sulfonylurea is associated with issues of hypoglycemia or weight gain, while SGLT2 inhibitor and GLP-1 RA have acceptable profiles with significant cardiovascular and renal benefit when used alone and with a weight loss effect in addition to insulin [2]. DPP-4 inhibitor use is rapidly and widely growing because it seems to be beneficial regarding glycemic control without significant adverse effects [8].

The proportion of patients with T1DM using basal analogs or rapid acting analogs increased, while the proportion of those patients using human insulin, BHI or premixed insulin decreased. In this study, the numbers of insulin users with T1DM were 34,972 in 2002, 46,011 in 2009, and 20,287 in 2019. The patients who were prescribed insulin and claiming both T1DM and T2DM were 5,117 in 2002, 8,959 in 2009, and 12,102 in 2019. The number of patients prescribed insulin for T1DM alone has decreased, and the number of patients prescribed insulin for both T1DM and T2DM has increased. In July 2011, the Korean NHIS promoted the T1DM registration so that registered patients could receive financial aid for consumable materials. After that, the patients with atypical T1DM which is defined as the T1DM with slowly progressive nature including latent autoimmune disease or long standing T2DM with insulin dependency increased in 2013, especially in age over 60 years [9]. Therefore, some insulin-using T2DM might be overestimated as T1DM. However, from around 2017, oral antidiabetic drug prescribed with the E10 ICD code are no longer covered by insurance, after that the proportion of T1DM might be underestimated. Song et al. [10] reported that the number of patients with T1DM from 2011 to 2013 using NHI data was 44,747 based on those who received insulin prescriptions three or more times. In the report, 85,764 patients were prescribed insulin at least once over the 2 years and 141,516 did not receive insulin prescription were among those claimed for T1DM [10]. In the Health Insurance Review & Assessment Service data, the numbers of patients with T1DM were 77,794 in 2017, 48,240 in 2019, and 44,827 in 2021 (opendata.hira.or.kr). These numbers were estimated through all claims claimed with primary ICD-10 code with E10 not confined to insulin claims. Therefore, the estimated numbers may have been overestimated when T1DM is defined by ICD-10 code alone. Therefore, in our study, we used a strict definition of T1DM that excluded patients who prescribed insulin during hospitalization and was limited to insulin users claiming T1DM alone to provide more accurate results. It is not easy to distinguish the types of diabetes in adults worldwide since T1DM has historically been characterized as having an onset during childhood [11], especially in Asia, the immunologic features and epidemiology of T1DM vary considerably compared to Caucasians, therefore, distinguishing type of diabetes even young patients is challenging [12]. In particular, it is estimated that the number of T1DM diagnoses could be affected by insurance policies in Korea. The purpose of this study was to analyze the trends in change of insulin only in cases where T1DM is certain (excluding group with overlapping T1DM and T2DM) rather than changes in the prevalence of T1DM [13].

According to diabetes factsheet published in 2022 by KDA, the number of patients being treated for T2DM was 3,900,596 in 2019. Among them, proportion of insulin users were 8.4%. In our study, the number of insulin user with T2DM was 348,254 in 2019. This number is close to the value reported in factsheet considering the definition of insulin users was more stringent in our study. In United States, 21.6% use insulin among patients with T2DM using antidiabetic medication in 2012 and in Japan, 17.9% use insulin in 2014 to 2016 [14,15]. The proportion of insulin user in T2DM is small compared to other counties. According to the reports regarding change in insulin prescription of T2DM in the United States, Japan, and Taiwan, analog insulin use largely has increased and premix insulin and human insulin use has decreased, similar to our study [16-19]. The proportion of long acting analog users was similar in each country: 65.5% in Japan (regardless of other insulin), 50% in United States (basal only), and 75% and 52% in Korea, regardless of other insulin and basal only. However, the proportion of basal plus bolus insulin use has remained constant in the United States, at about 40% among insulin users [18], while in Korea, the proportion using basal plus bolus insulin has increased steadily in T2DM but reaches only 17%. In contrast, among the patients using insulin with T2DM, the proportion of patients taking insulin only or insulin plus non-insulin monotherapy has decreased since 2014; the proportion of insulin plus non-insulin dual therapy has increased markedly since 2015 and has reached over 50% since 2018. This trend is similar to other studies reported in Japan [17] and Taiwan [16]. These studies reported that the patients using non-insulin drug in combination with insulin is much more prevalent than the patients using insulin alone, and the basal insulin only group or basal plus bolus group prefers to use non-insulin drug in combination with insulin than bolus only group. This is because combining other hypoglycemic agents with basal insulin is easier for patients to take and easier for physicians to educate than to intensify insulin therapy such as adding bolus insulin to basal insulin. In addition, intensified insulin therapy without proper education, multiple self-monitoring of glucose, or continuous glucose monitoring use might be associated with hypoglycemia, cardiovascular risk, and mortality in both T1DM and T2DM patients [20-23]. Considering that diabetes is progressive disease, patients with T2DM who need to use insulin might have decreased residual β-cell function. Without effective education on insulin dose control, continuous daily insulin dose titration by patients themselves is not effective. Additionally, inappropriate insulin dose adjustment could induce hypoglycemia and increase glycemic variability, which are associated with increased cardiovascular risk. Therefore, from a long-term perspective, an education system in the hospital supported by government policies that can support appropriate and intensified insulin administration when insulin is needed is essential for the treatment of patients with DM and decreased residual β-cell function.

This study evaluated the insulin trends among insulin users with T1DM and T2DM. We also investigated the trends of antidiabetic drug use in T2DM insulin users from 2002 to 2019. A strength of this study is the use of the NHIS data, which encompasses almost all patients prescribed insulin in Korea. Therefore, we were able to assess the status of insulin use in almost the entire Korean population. A limitation of this study is that the diabetes classification is based on the ICD-10 code and outpatient insulin prescriptions; thus numbers for T1DM or T2DM patients may be overestimated or underestimated. In addition, we could not identify HbA1c and the status of glycemic control because the NHIS data set does not included this information. In the future, studies for the glycemic control status according to insulin type in insulin users with T1DM and T2DM or according to antidiabetic medication in insulin users with T2DM are needed.

In conclusion, this study identified important patterns of insulin use in Korea in T1DM and T2DM patients, including trends of predominant use of insulin and intensified insulin regimen. We also evaluated the antidiabetic medication trends in T2DM. Such knowledge will be informative for not only clinicians who treat diabetes but also health policy managers to manage diabetes in Korea.