Correlation between diabetes complications severity index and actual complication severity and diabetic medication in Japanese patients with type 2 diabetes mellitus

The Diabetes Complications Severity Index (DCSI) is scored from 0 to 2 points for each condition based on the progression of diabetic polyneuropathy (DPN), diabetic retinopathy (DR), diabetic nephropathy (DN), cardiovascular disease (CVD), stroke, and peripheral arterial disease (PAD). It has been reported that the DCSI score is a good indicator of the association between diabetic complications, mortality, and health care costs.1-5 However, there are no prior reports on the usefulness of DCSI in Japanese individuals or its association with the future initiation of insulin therapy. The purpose of this study is to investigate the correlation between DCSI and diabetes treatment at discharge in Japanese hospitalised patients with type 2 diabetes mellitus (T2D). The subject flowchart for this study is shown in Supplemental Figure 1. Ultimately, 553 patients with T2D who underwent educational hospitalisation6 in the Department of Diabetes, Endocrinology, and Metabolism at Kawasaki Medical School Hospital between 1 April 2018, and 31 December, 2023, were included in this study. Supplemental Table 1 shows the definition of the DCSI score used in this study. This definition of the DCSI score was adapted from previous reports on the DCSI,1 incorporating the definition of complications we used in our past research.7 DCSI scoring was independently evaluated by two researchers while reviewing actual clinical data and medical records. When evaluations agreed that the score was used for analysis, no consultation was conducted. However, when evaluations differed, consultation was conducted to determine the final score. This study is a single-centre retrospective study. All data for analysis were collected from medical records. Participants in this study generally received multiple daily insulin injections during hospitalisation, and other diabetes medications were discontinued during this period. Hospital admission medical records, endogenous insulin secretion capacity indicators, and tests for each complication were performed according to our previously published method.8 Data were expressed as median (interquartile range). The primary aim of this study is to clarify the correlation between DCSI and treatment at discharge in Japanese patients with T2D admitted to our hospital. The secondary aim is to evaluate whether DCSI correlates with the progression of chronic diabetic complications in Japanese patients with T2D. For the comparison of clinical parameters within each group, the Kruskal–Wallis test was used. Post hoc multiple comparisons were performed using the Steel–Dwass test to control for family-wise error within each variable. For categorical variables, the chi-square test was used. To identify factors associated with insulin therapy at discharge, we performed logistic regression analysis. The independent variables included DCSI plus factors previously identified in our reports.8 In this study, p <0.05 was set as the statistical significance level. Statistical analysis was performed using JMP Student Edition 18, and figures were created using Microsoft Excel for Mac (16.100). The study protocol (including opt-out informed consent) was approved by the Kawasaki Medical School Ethics Committee (Approval No. 6362-02). This study was conducted in accordance with the Declaration of Helsinki. Participants were grouped as follows, based on the DCSI score quartile range: Lowest group (DCSI 0 points, n = 143), Low group (DCSI 1 point, n = 157), High group (DCSI 2–3 points, n = 195), Highest group (DCSI ≥4 points, n = 77). Table 1 shows clinical characteristics. There were no differences in blood glucose levels, HbA1c, or C-peptide index (CPI) at admission among the groups. Age and duration of T2DM were longer with higher DCSI scores. Higher DCSI scores were associated with better outcomes for lipid parameters. The use of medication for hypertension, lipid-lowering drugs, including statins, was higher in the group with higher DCSI scores (Supplementary Figure 2). These findings suggest that patients with higher DCSI scores tend to be older with longer diabetes duration and better-controlled lipid profiles, suggesting more intensive chronic care. Figure 1 shows the prevalence of diabetic complications in each group classified by DCSI. The prevalence of BDC-based DPN, DR, and DN in NCS was higher in participants with higher DCSI (Figure 1A–C). Vibration persistence time, an indicator of DPN, and the coefficient of variation of R-R intervals during deep breathing (CVR-R), an indicator of diabetic autonomic neuropathy, were significantly reduced in the group with higher DCSI (Table 1). Objective measures of diabetic neuropathy declined with increasing DCSI scores, supporting the utility of DCSI in reflecting not only self-reported or physician-diagnosed complications but also quantifiable physiological dysfunction. The proportion of participants with a history of CVD, stroke, or foot ulcers was higher in the group with higher DCSI scores (Figure 1D). The mean and maximum carotid intima-media thickness, a marker of diabetic macrovascular disease progression, and ankle brachial index also showed the same trend. The utilisation rates of diabetes medications at admission or discharge are shown in Figure 1E,F. The rate of insulin therapy was higher at discharge compared to admission in all groups except the lowest group (p <0.001). The usage rates of glucagon-like peptide-1 receptor agonists and sulfonylurea/glinide were higher in groups with higher DCSI scores. The usage rate of biguanide was higher in groups with lower DCSI scores. Finally, logistic regression analysis was performed to identify factors associated with insulin therapy at discharge (Supplementary Table 2). DCSI was an independent factor related to the use of insulin therapy at discharge, in addition to known factors. This study examined the clinical significance of the DCSI, a comprehensive complication indicator, in Japanese patients with T2D undergoing educational hospitalisation. Our findings demonstrate a significant correlation between DCSI and treatment intensity at discharge, suggesting its utility as a practical indicator for decision-making in clinical practice. Most prior studies on DCSI have focused on long-term outcomes in outpatient settings,2, 4, 9, 10 and large cohort studies of hospitalised patients included many non-diabetic patients with heart failure.11 Our study is unique in assessing its association with treatment decisions and objective complication markers during hospitalisation. We provide the first evidence that DCSI is quantitatively associated with markers for DPN and atherosclerosis, such as BDC, CVR-R, and carotid IMT. DCSI was a clinically meaningful and easily applicable index, reflecting both treatment intensity at discharge and the severity of diabetes-related complications. The higher rates of insulin use and GLP-1 receptor agonist therapy in the higher DCSI groups may reflect clinicians' tendency to intensify treatment in patients with a higher burden of complications, particularly in the context of educational hospitalisation where individualised care plans are emphasised. Conversely, the lower use of metformin in higher DCSI groups may be related to renal dysfunction, a component of the DCSI, and concerns about lactic acidosis. This study has several limitations. First, this was a retrospective single-centre study in a specific context of educational hospitalisation, which may limit generalizability. Second, insulin initiation at discharge may have been influenced by institutional protocols and physician discretion.6, 12 Third, the study period overlapped with the COVID-19 pandemic, but infection history was not considered (Supplemental Table 3, Supplemental Figure 3). This limitation may be partly influenced by unmeasured COVID-19-related factors. In addition, DCSI includes subjective components based on medical history or clinical judgement, potentially introducing variability. Finally, while we observed statistical associations, the cross-sectional design precludes any inference of causality. Therefore, the current findings should be interpreted as hypothesis-generating. Taken together, DCSI may serve as a simple and clinically meaningful index for assessing complication burden and guiding treatment intensity in Japanese patients with T2D. While it could enhance risk stratification in routine care, further multicenter studies across diverse populations are needed to validate and refine its application. Y.I. researched data, analyzed all data, and wrote the manuscript. Y.I. and H.I. independently performed DCSI scoring for all patients. H.I., T.K., M.K., Y.O., R.I., T.I., T.S., K.D., Y.F., J.S., M.S., S.N., K.K. participated in the discussion. H.K. participated in the discussion and reviewed the manuscript. There is no funding associated with this manuscript. H.K. has received honoraria for lectures, received scholarship grants, and received research grants from Novo Nordisk Pharma, Sanofi, Eli Lilly, Boehringer Ingelheim, Taisho Pharma, Sumitomo Pharma, K.K. has been an advisor to, received honoraria for lectures from, and received scholarship grants from Novo Nordisk Pharma, Sanwa Kagaku, Taisho Pharma, Sumitomo Pharma, Mitsubishi Tanabe Pharma, Boehringer Ingelheim. S.N. has received honoraria for lectures from Novo Nordisk Pharma and Daiichi Sankyo. T.K. has received honoraria for lectures from Novo Nordisk Pharma and Sumitomo Pharma. All other authors have no conflicts of interest. The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/dom.70321. All data generated or analyzed during this study are included in this published article. Supplemental Figure 1. Flowchart of participants with type 2 diabetes mellitus in this study. Supplemental Figure 2. Use of medication for hypertension or dyslipidemia on admission Supplemental Figure 3. Use of diabetes medication before and after educational hospitalisation across DCSI score quartile groups. Supplemental Table 1. Calculation Method for the Diabetes Complication Severity Index Score Supplemental Table 2. Logistic regression analysis to identify factors associated with the use of insulin therapy at discharge from educational hospitalisation. Supplemental Table 3. Differences in clinical characteristics in this study beefore and after the Coronavirus Disease 2019 (COVID-19) Pandemic. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.