Factor affecting severe atherothrombotic cerebral infarction in patients with type 2 diabetes mellitus: Large‐scale claim database analysis of Japan

ABSTRACT

Aims

This study aimed to investigate the factors associated with the exacerbation of the severity of atherothrombotic brain infarction at discharge in patients with type 2 diabetes using a large‐scale claims database.

Materials and Methods

This retrospective cross‐sectional study utilized the Medical Data Vision administrative claims database, a nationwide database in Japan using acute care hospital data, and the Diagnosis Procedure Combination system. Diagnosis Procedure Combination data collected between April 1, 2008, and December 31, 2022, were extracted. Patients with type 2 diabetes were included. Severe atherothrombotic brain infarction was defined as a modified Rankin scale score of ≥3.

Results

Severe atherothrombotic brain infarction occurred in 43,916/99,864 (44.0%) patients with type 2 diabetes. The odds ratio for severe atherothrombotic brain infarction increased significantly per 10 year increments in age (odds ratio: 1.69, 95% confidence interval: 1.66–1.71). A body mass index of <25 kg/m², with a body mass index of ≥25 kg/m² as reference, also increased the risk for severe atherothrombotic brain infarction (odds ratio: 1.11, 95% confidence interval: 1.08–1.15). The odds ratios in insulin and dipeptidyl peptidase 4 inhibitor use were significantly higher than 1. In particular, statin use (odds ratio: 0.85, 95% confidence interval: 0.83–0.88), fibrate use (odds ratio: 0.68, 95% confidence interval: 0.59–0.78), aspirin use (odds ratio: 0.78, 95% confidence interval: 0.75–0.80), and P2Y12 inhibitor use (odds ratio: 0.88, 95% confidence interval: 0.85–0.91) were associated with a lower odds ratio for severe atherothrombotic brain infarction.

Conclusions

The active management of lipid levels using statins and fibrates may be beneficial in preventing the exacerbation of atherothrombotic brain infarction in type 2 diabetes patients.

The active management of lipid levels using statins and fibrates may be beneficial in preventing the exacerbation of atherothrombotic brain infarction.

INTRODUCTION

The annual incidence of cardiovascular events, such as ischemic heart disease and stroke, has been increasing in Japan. Many of these events are caused by thrombosis, which occurs due to plaque rupture and is collectively referred to as atherothrombosis. This group of diseases accounts for approximately one‐quarter of deaths in Japan, making it an urgent challenge to understand its pathogenesis and to establish effective prevention and treatment methods.

While Western populations have a high incidence of myocardial infarction, the Japanese population has a high incidence of stroke. This ethnic difference has been noted in studies such as the Seven Country Study ¹ and the Nihon‐San Study ² . Although this difference has decreased in recent years, studies such as the Hisayama Study ³ have shown that the incidence of stroke still exceeds that of myocardial infarction. East Asians, including the Japanese, carry a significant genetic susceptibility variant for cardiovascular diseases, namely, the RNF213p.R4810K polymorphism ⁴ , ⁵ , ⁶ . This genetic variant is specific to East Asians and has not been previously identified in Europeans. The RNF213p.R4810K polymorphism is associated with a 3.58‐fold increase in the risk of developing atherothrombotic stroke. Stroke is currently the fourth leading cause of death in Japan, and it requires attention owing to its significant impact on the quality of life of patients with post‐stroke disabilities.

The incidence of atherothrombotic brain infarction (ATBI), characterized by arteriosclerosis, is increased across all types of strokes. Diabetes mellitus induces inflammation in blood vessels due to sustained hyperglycemia, leading to vascular damage and the promotion of vascular wall remodeling and atheroma formation. Numerous cohort studies and meta‐analyses ⁷ , ⁸ have established diabetes mellitus as a major risk factor for stroke. In the Stroke Data Bank 2021, 33.9% of the patients who developed atherothrombotic cerebral infarction also had diabetes mellitus ¹ . Considering that hypertension and dyslipidemia are closely associated with the risk factors of atherothrombotic cerebral infarction, they are important complications to be aware of in patients with diabetes mellitus.

The treatment for ATBI is aimed to maintain the pre‐onset level of activity and activities of daily living while preventing recurrence. A report from the Japan Stroke Data Bank ⁹ indicates that patients without diabetes mellitus have a higher rate of favorable prognosis at discharge following hospitalization for atherothrombotic cerebral infarction. Furthermore, patients with diabetes mellitus had an odds ratio (OR) of 1.29 for poor outcomes defined as a modified Rankin scale (mRS) score of ≥3 ⁹ . As the onset of atherothrombotic cerebral infarction significantly affects the prognosis and post‐onset quality of life of patients with diabetes, numerous pharmacological therapies and intervention studies aimed at suppressing onset have been developed. GLP‐1 receptor agonists and pioglitazone, which are blood glucose‐lowering drugs available in Japan, have been reported to have anti‐atherosclerotic effects ¹⁰ , ¹¹ , ¹² . Additionally, statins and fibrates, drugs used for the treatment of chronic diseases, have been reported to have inhibitory effects on atherosclerosis ¹³ , ¹⁴ . Therefore, a good quality of life can still be maintained even if a patient develops atherothrombotic cerebral infarction if the functional impairment is mild. Therapies that demonstrate inhibitory effects on atherosclerosis may suppress the severity of atherothrombotic cerebral infarction; however, detailed studies on this topic are extremely limited. Thus, in this study, we aimed to investigate the factors that could suppress the severity of atherothrombotic cerebral infarction at discharge in patients with type 2 diabetes, using a large‐scale claims database.

MATERIALS AND METHODS

Study design

This retrospective cross‐sectional study was conducted using the Medical Data Vision (MDV) administrative claims database, a nationwide database in Japan using acute care hospital data, and the Diagnosis Procedure Combination (DPC) system. Briefly, the MDV database is a hospital‐based claims database covering approximately 45 million cumulative patients. As of December 2023, it included patients who received inpatient or outpatient treatment at acute care hospitals (28% of all hospitals) using 493 DPC systems in Japan. The DPC system is a healthcare payment system used in Japan. It is designed to control medical costs and to standardize hospital charges by classifying patient diagnoses and treatments into specific groups.

Approximately 4.7 million (10%) of the patients were diagnosed with diabetes mellitus. The MDV database contains information on patient characteristics, such as age and sex, diagnosed diseases, medical procedures performed, and prescribed medications. All patient data are encrypted by the MDV when incorporated into the database, ensuring that patient information is de‐identified for all users of the MDV database. The current study analyzed DPC data collected between April 1, 2008, and December 31, 2022. Patients diagnosed with ATBI for the first time and having data related to subsequent occurrences were excluded. Prescribed medications and targeted drugs used up to the day of admission for ATBI were investigated.

Study population

First, the data of 4,702,688 patients with diabetes mellitus (International Classification of Diseases, 10th revision [ICD‐10] code: E11–E14) registered between April 1, 2008, and December 31, 2022, in the MDV claim database were extracted (Figure 1). Next, patients with type 2 diabetes who had been prescribed hypoglycemic agents were identified. The following patients were excluded: (i) patients without type 2 diabetes (n = 1,487,591); (ii) patients diagnosed with type 2 diabetes and type 1 diabetes or any diabetes mellitus (n = 317,727); (iii) patients with no prescription for hypoglycemic agents (n = 1,204,490); (iv) patients without atherothrombotic stroke (n = 1,593,015); (v) patients without brain imaging (computed tomography [CT]/magnetic resonance imaging [MRI]) data (n = 1,324); and (vi) patients with a missing mRS score at discharge (n = 1,336). Finally, 99,864 patients with type 2 diabetes and ATBI were included and subsequently divided into two groups: (i) patients with a modified mRS score of <3 at discharge (n = 55,949) and (ii) patients with an mRS score of ≥3 (severe ATBI) at discharge (n = 43,916). In the sub‐analysis, patients without the laboratory data under investigation were excluded, while (i) those with mRS <3 at discharge (n = 6,214) and (ii) those with an mRS of ≥3 at discharge (n = 2,717) were evaluated. Diseases defined solely by ICD‐10 codes have been reported to have low specificity ¹⁵ . Therefore, this study identified the patients with type 2 diabetes as those prescribed with hypoglycemic agents for type 2 diabetes.

Figure 1.

Patient disposition. mRS, modified Rankin scale; T1D, type 1 diabetes; T2D, type 2 diabetes.

Identification of ATBI events and use of concomitant drugs

ATBI events were identified based on an ICD‐10 code of I633 assigned at the time of disease diagnosis and based on findings on CT or MRI scan performed within 3 days before or after admission. In this study, atorvastatin, pitavastatin, and rosuvastatin were classified as strong statins, whereas other statins were classified as standard statins.

Outcomes

The severity of ATBI was assessed using the mRS. The mRS is a commonly used scale for measuring the degree of disability in the daily activities of people who have experienced a stroke ¹⁶ , ¹⁷ . For outcome prediction in clinical trials, the mRS is usually dichotomized, where a good functional outcome is reflected by a score of 0–2 and a poor functional outcome by a score of ≥3 ¹⁸ . In the current study, severe ATBI was defined as an mRS score of ≥3 at discharge.

Patient characteristics

Age, sex, and body mass index (BMI) were identified using data documented within 30 days of admission in the claims records. Obesity was defined as a BMI of ≥25 kg/m². Laboratory data were collected to determine the estimated glomerular filtration rate (eGFR), hemoglobin A1c (HbA1c), high‐density lipoprotein‐cholesterol (HDL‐C), low‐density lipoprotein‐cholesterol (LDL‐C), and triglyceride (TG). Cutoffs were set at <7.0% for HbA1c; ≥40 mg/dL, HDL; <120 mg/dL, LDL; <220 mg/dL, TG; and <60 mL/min/1.73 m², eGFR according to the guidelines ¹⁹ .

Statistical analysis

Normally distributed data (age, BMI, eGFR, HbA1c, HDL‐C, LDL‐C, and TG) were expressed as mean ± SD and analyzed using an unpaired t‐test. Meanwhile, categorical variables were expressed as absolute numbers or percentages and analyzed using the χ²‐test or Fisher's exact test. Factors significantly associated with severe ATBI were identified using univariate analysis, and significant factors (i.e., with P < 0.2 in the univariate analysis) were included in the multivariate analysis. Risks were expressed as ORs with their 95% confidence intervals (CIs). All statistical analyses were performed using the IBM SPSS Statistics for Windows (version 23.0; IBM Corp., Armonk, NY, USA). Statistical significance was set at P < 0.05.

RESULTS

Table 1 presents the baseline patient characteristics. The mean patient age and BMI were 74.2 ± 11.1 years and 23.7 ± 4.0 kg/m², respectively. Severe ATBI was observed in 43,916 of the 99,864 patients (44.0%). There were significant differences in many clinical parameters between the severe and non‐severe ATBI groups. Compared with the non‐severe ATBI group, the severe ATBI group involved a larger proportion of female patients and patients with an mRS score of ≥ 3 on preadmission and were also older and had a lower BMI. The frequency of hypoglycemic, antihypertensive, antihyperlipidemic, and antiplatelet drug use in each group is shown in Table 1.

Table 1.

Baseline characteristics following atherothrombotic ischemic stroke onset in patients with type 2 diabetes

	Characteristics	Overall (n = 99,864)	mRS score <3 (n = 55,948)	mRS score ≥3 (n = 43,916)	P‐value
Sex, n (%)	Male	65,293 (65.4)	39,669 (70.9)	25,624 (58.3)	<0.001
	Female	34,571 (34.6)	16,279 (29.1)	18,292 (41.7)
Age	Mean (years)	74.2 ± 11.1	71.4 ± 10.9	77.8 ± 10.3	<0.001
BMI	Mean (kg/m2)	23.7 ± 4.0	24.2 ± 3.9	23.1 ± 4.0	<0.001
	Distribution, n (%)
	<25	66,920 (67.0)	35,386 (63.2)	31,534 (71.8)	<0.001
	≥25	32,944 (23.0)	20,562 (36.8)	12,382 (28.2)
mRS score before atherothrombotic stroke onset	Distribution, n (%)
	0	47,046 (47.1)	31,768 (56.8)	15,278 (34.8)	<0.001
	1	19,154 (19.2)	12,288 (22.0)	6,866 (15.6)
	2	12,779 (12.8)	7670 (13.7)	5,109 (11.6)
	3	9,089 (9.1)	2,408 (4.3)	6,681 (15.2)
	4	8,248 (8.3)	1,372 (2.5)	6,876 (15.7)
	5	2,990 (3.0)	217 (0.4)	2,773 (6.3)
	6	0 (0)	0 (0)	0 (0)
mRS score at discharge	Distribution, n (%)		2548 (5.9)	3326 (5.8)	<0.001
	0	10,440 (10.5)	10,440 (18.7)	0 (0)	<0.001
	1	25,238 (25.3)	25,238 (45.1)	0 (0)
	2	20,270 (20.3)	20,270 (36.2)	0 (0)
	3	14,282 (14.3)	0 (0)	14,282 (32.5)
	4	17,967 (18.0)	0 (0)	1,7967 (40.9)
	5	8,147 (8.2)	0 (0)	8,147 (18.6)
	6	3,520 (0.04)	0 (0)	3,520 (8.0)
mRS score <3 at discharge	Distribution, n (%)
	Yes	55,948 (56.0)	55,948 (100.0)	0 (0)	<0.001
	No	43,916 (44.0)	0 (0)	43,916 (100.0)
Tissue plasminogen activator	Distribution, n (%)	5,874 (5.9)	2,548 (4.5)	3,326 (7.6)	<0.001
Anti‐diabetic agent use	Distribution, n (%)
DPP‐4 inhibitors		19,805 (19.8)	10336 (18.5)	9,469 (21.6)	<0.001
Glinides		2,442 (2.4)	1,218 (2.2)	1,224 (2.8)	<0.001
GLP‐1 receptor agonists		526 (0.5)	257 (0.5)	269 (0.6)	0.009
Insulin		19,373 (19.4)	8,964 (16.0)	10,409 (23.7)	<0.001
Metformin		8115 (8.1)	4,818 (8.6)	3,297 (7.5)	<0.001
SGLT2is		3,329 (3.3)	1,955 (3.5)	1,374 (3.1)	0.014
Sulfonylureas		6,798 (6.8)	3,628 (6.5)	3,170 (7.2)	<0.001
Thiazolidinediones		1,848 (1.9)	1,060 (1.9)	788 (1.8)	0.243
α‐GIs		5486 (5.5)	2,927 (5.2)	2,559 (5.8)	<0.001
Antiplatelet/anticoagulant drug use	Distribution, n (%)
Aspirin		32,463 (32.5)	20,345 (36.4)	12,118 (27.6)	<0.001
P2Y12 inhibitors		27,004 (27.0)	16,826 (30.1)	10,178 (23.2)	<0.001
Antihypertensive drug use	Distribution, n (%)
ACE inhibitors		2,545 (2.5)	1331 (2.4)	1,214 (2.8)	0.29
Aldosterone receptor blocker use		2,520 (2.5)	1,066 (1.9)	1,454 (3.3)	<0.001
ARBs		197 (0.2)	103 (0.2)	94 (0.2)	<0.001
Calcium channel blockers		18,957 (19.0)	10,624 (19.0)	8,333 (19.0)	0.955
Loop diuretics		6,004 (6.0)	2,477 (4.4)	3,527 (8.0)	<0.001
Thiazide diuretics		1,668 (1.7)	916 (1.6)	752 (1.7)	0.358
Tolvaptan		623 (0.6)	216 (0.4)	407 (0.9)
α‐Receptor blockers		1,471 (1.5)	723 (1.3)	748 (1.7)	<0.001
β‐Receptor blockers		8,622 (8.6)	4,517 (8.1)	4,105 (9.3)	<0.001
Cholesterol‐lowering drug use	Distribution, n (%)
Statins		23,569 (23.6)	14,534 (26.0)	9,035 (20.6)	<0.001
Fibrates		1,038 (1.0)	712 (1.3)	326 (0.7)	<0.001

Data are presented as number (%) or the mean (SD). The P‐value is calculated for the difference between patients with atherothrombotic ischemic stroke who have an mRS score of <3 and ≥3 at discharge. ACE inhibitors, angiotensin‐converting enzyme inhibitors; ARBs, angiotensin II receptor blockers; BMI, body‐mass index; DPP‐4 inhibitors, dipeptidyl peptidase‐4 inhibitors; GLP‐1 receptor agonists, glucagon‐like peptide‐1 agonists; mRS, modified Rankin scale; SGLT2is, sodium–glucose cotransporter 2 inhibitors; α‐GI, alpha‐glucosidase inhibitor.

The risk of severe ATBI is shown in Table 2. Compared with the male sex, the female sex was associated with a higher OR for severe ATBI (OR: 1.35, 95% CI: 1.31–1.39). The OR for severe ATBI significantly increased per 10 year increments in age (OR: 1.69, 95% CI: 1.66–1.71). A BMI of <25 kg/m², with BMI ≥25 kg/m² as the reference, also showed significance (OR: 1.11, 95% CI: 1.08–1.15). The ORs for insulin and dipeptidyl peptidase 4 inhibitor (DPP‐4I) use were significantly higher than 1. In particular, the uses of statins, fibrates, aspirin, and P2Y12 inhibitors were associated with lower ORs for severe ATBI. Several statins and fibrates were investigated in detail (Tables S1 and S2). In this study, statin and fibrate administration rates were 23.6% and 1.0%, respectively. The ORs for strong and standard statins showed similar trends. Fibrate use was associated with lower ORs for severe ATBI. Among fibrates, pemafibrate, which exerted anti‐peroxisome proliferator‐activated receptor alpha (PPARα) action, was associated with lower ORs for severe ATBI. Owing to the limited number of patients who received clinofibrate, it was not possible to analyze its effects. Among the antihypertensive agents used, the use of loop diuretics, tolvaptan, and aldosterone receptor antagonists was associated with higher ORs for severe ATBI.

Table 2.

Logistic regression analysis for risk factors of severe ATBI at discharge in patients with type 2 diabetes

Characteristic	Odds ratio	95% CI	P‐value
Sex
Male	Reference	–	–
Female	1.35	1.32–1.39	<0.001
Age
/10 year increase	1.69	1.66–1.71	<0.001
BMI
≥25	Reference	–	–
<25	1.11	1.08–1.15
Tissue plasminogen activator use	1.61	1.52–1.70	<0.001
Anti‐diabetic agent use
DPP‐4 inhibitors	1.16	1.12–1.21	<0.001
Glinides	1.04	0.95–1.14	0.43
GLP‐1 receptor agonists	1.18	0.98–1.42	0.08
Insulin	1.75	1.69–1.81	<0.001
Metformin	0.96	0.91–1.02	0.18
SGLT2is	0.93	0.86–1.00	0.06
Sulfonylureas	0.98	0.92–1.03	0.40
Thiazolidinediones	–	–	–
α‐GIs	0.97	0.91–1.03	0.32
Antiplatelet/anticoagulant drug use
Aspirin	0.80	0.77–0.82	<0.001
P2Y12 inhibitors	0.89	0.87–0.92	<0.001
Antihypertensive drug use
ACE inhibitors	1.04	0.95–1.14	0.38
Aldosterone receptor blocker use	1.19	1.08–1.31	<0.001
ARBs	0.93	0.89–0.93	<0.001
Calcium channel blockers	–	–	–
Loop diuretics	1.40	1.31–1.50	<0.001
Thiazide diuretics	–	–	–
Tolvaptan	1.31	1.09–1.57	<0.001
α‐Receptor blockers	1.21	1.08–0.88	<0.001
β‐Receptor blockers	1.04	0.95–1.14	0.18
Cholesterol‐lowering drug use
Statins	0.86	0.83–0.89	<0.001
Fibrates	0.68	0.60–0.78	<0.001

The data show the results of the multivariate analysis. Thiazolidinediones, calcium channel blockers, and thiazide diuretics are excluded from the multivariate analysis because they are not significant (i.e., P > 0.2) in the univariate analysis. ACE inhibitors, angiotensin‐converting enzyme inhibitors; ARBs, angiotensin II receptor blockers; BMI, body mass index; CI, confidence interval; DPP‐4 inhibitors, dipeptidyl peptidase‐4 inhibitors; GLP‐1 receptor agonists, glucagon‐like peptide‐1 agonists; SGLT2is, sodium‐glucose cotransporter 2 inhibitors; α‐GI, alpha‐glucosidase inhibitor.

The risk of severe ATBI in patients with available laboratory data is shown in Tables 3 and 4. The ORs for sex, age, BMI, and concomitant medications were similar to those in the analysis that did not include laboratory data. The value was significant for an eGFR of <30 mL/min/1.73 m² (OR: 1.19, 95% CI: 1.04–1.35) with eGFR of ≥60 mL/min/1.73 m² as the reference. HbA1c <7.0% and HDL‐C >40 mg/dL were associated with a significantly lower OR for severe ATBI when the target set by the guidelines was achieved. However, the OR for severe ATBI significantly increased when TG was <150 mg/dL achieved their targets set by the guidelines.

Table 3.

Baseline characteristics of the patients with type 2 diabetes with available laboratory values following atherothrombotic ischemic stroke onset

	Characteristics	Overall (n = 8,931)	mRS score <3 (n = 4,904)	mRS score ≥3 (n = 4,027)	P‐value
Sex, n (%)	Male	5,819 (65.2)	3,462 (70.6)	2,357 (58.5)	<0.001
	Female	3,112 (41.5)	1,442 (29.4)	1,670 (41.5)
Age	Mean (years)	74.2 ± 11.0	71.4 ± 10.8	77.6 ± 10.4	<0.001
BMI	Mean (kg/m2)	23.8 ± 4.1	24.3 ± 4.0	23.2 ± 4.1	<0.001
	Distribution, n (%)
	<25	5,932 (66.4)	3,074 (62.7)	2,858 (71.0)	<0.001
	≥25	2,999 (33.6)	1,830 (37.3)	1,169 (29.0)
eGFR	Mean (mL/min/1.73 m2)	47.7 ± 23.0	49.5 ± 22.2	45.5 ± 23.8	<0.001
	Distribution, n (%)
	≥60	2,710 (30.3)	1,614 (32.9)	1096 (27.2)	<0.001
	31–59	4,188 (46.9)	2,355 (48.0)	1833 (45.5)
	<30	2,033 (22.8)	935 (19.1)	1098 (27.3)
HbA1c	Mean (%)	7.6 ± 1.8	7.6 ± 1.8	7.7 ± 1.8	0.098
	Distribution, n (%)
	≥7.0	5,021 (56.2)	2,719 (55.4)	2,302 (57.2)	0.103
	<7.0	3,910 (43.8)	2,185 (44.6)	1,725 (42.8)
HDL‐C	Mean (mg/dL)	57.5 ± 16.9	58.6 ± 17.1	56.2 ± 16.5	<0.001
	Distribution, n (%)
	<40	1,025 (11.5)	479 (9.8)	546 (13.6)
	≥40	7,906 (88.5)	4,425 (90.2)	3,481 (86.4)
LDL‐C	Mean (mg/dL)	125.4 ± 39.9	126.3 ± 39.4	124.4 ± 40.5	0.087
	Distribution, n (%)
	<120	4,240 (47.5)	2,273 (46.3)	1,967 (48.8)	0.019
	≥120	4,691 (52.5)	2,631 (53.7)	2,060 (51.2)
TG	Mean (mg/dL)	187.3 ± 147.2	200.3 ± 161.0	171.5 ± 126.7	<0.001
	Distribution, n (%)
	<150	4,356 (51.2)	2,164 (44.1)	2,192 (45.6)	<0.001
	≥150	4,575 (48.8)	2,740 (55.9)	1,835 (54.4)
mRS score before atherothrombotic stroke onset	Distribution, n (%)
	0	4,206 (47.1)	2,794 (57)	1,412 (35.1)	<0.001
	1	1,787 (20)	1,083 (22.1)	704 (17.5)
	2	1,149 (12.9)	671 (13.7)	478 (11.9)
	3	832 (9.3)	204 (4.2)	628 (15.6)
	4	689 (7.7)	125 (2.5)	564 (14)
	5	237 (2.7)	16 (0.3)	221 (5.5)
	6	0 (0)	0 (0)	0 (0)
mRS score <3 at discharge	Distribution, n (%)
	Yes	4,904 (54.9)	4,904 (100)	0 (0)	<0.001
	No	4,027 (45.1)	0 (0)	4,027 (100)
mRS score at discharge	Distribution, n (%)
	0	863 (9.7)	863 (17.6)	0 (0)	<0.001
	1	2,196 (24.6)	2,196 (44.8)	0 (0)
	2	1,845 (20.7)	1,845 (37.6)	0 (0)
	3	1,310 (14.7)	0 (0)	1,310 (32.5)
	4	1,675 (18.8)	0 (0)	1,675 (41.6)
	5	755 (8.5)	0 (0)	755 (18.7)
	6	287 (3.2)	0 (0)	287 (7.1)
Tissue plasminogen activator use	Distribution, n (%)	252 (2.8)	252 (5.1)	0 (0)	<0.001
Anti‐diabetic agent use	Distribution, n (%)
DPP‐4 inhibitors		2,085 (23.3)	1,037 (21.1)	1,048 (26.0)	<0.001
Glinides		240 (2.7)	119 (2.4)	121 (3.0)	0.093
GLP‐1 receptor agonists		58 (0.6)	25 (0.5)	33 (0.8)	0.07
Insulin		1,902 (21.3)	823 (16.8)	1,079 (26.8)	<0.001
Metformin		839 (9.4)	493 (10.1)	346 (8.6)	<0.001
SGLT2is		438 (4.9)	257 (5.2)	181 (4.5)	0.104
Sulfonylureas		698 (7.8)	339 (6.9)	359 (8.9)	<0.001
Thiazolidinediones		221 (2.5)	117 (2.4)	104 (2.6)	0.551
α‐GIs		510 (5.7)	275 (5.6)	235 (5.8)	0.644
Antiplatelet/anticoagulant drug use	Distribution, n (%)
Aspirin		2,767 (31.0)	1,640 (33.4)	1,127 (28.0)	<0.001
P2Y12 inhibitors		2,415 (27.0)	1,447 (29.5)	968 (24.0)	<0.001
Antihypertensive drug use
ACE inhibitors		336 (3.8)	179 (3.7)	157 (3.9)	0.539
Aldosterone receptor blocker use		269 (3.0)	113 (2.3)	156 (3.9)	<0.001
ARBs		1,538 (17.2)	857 (17.5)	681 (16.9)	0.482
Calcium channel blocker		2,048 (22.9)	1,127 (23.0)	921 (22.9)	0.276
Loop diuretics		620 (6.9)	249 (5.1)	371 (9.2)	<0.001
Thiazide diuretics		192 (2.1)	98 (2.0)	94 (2.3)	0.276
Tolvaptan		63 (0.7)	21 (0.4)	42 (1.0)	<0.001
α‐receptor blockers		165 (1.8)	92 (1.9)	73 (1.8)	0.825
β‐receptor blockers		886 (9.9)	452 (9.2)	434 (10.8)	0.014
Cholesterol‐lowering drug use	Distribution, n (%)
Statins		2,354 (26.4)	1,407 (28.7)	947 (23.5)	<0.001
Fibrates		112 (1.3)	67 (1.4)	45 (1.1)	0.293

Data are presented as number (%) or the mean (SD). P‐values are calculated for the difference between patients with atherothrombotic ischemic stroke who have an mRS score of <3 and ≥3 at discharge. ACE inhibitors, angiotensin‐converting enzyme inhibitors; ARBs, angiotensin II receptor blockers; BMI, body mass index; DPP‐4 inhibitors, dipeptidyl peptidase‐4 inhibitors; eGFR, estimated glomerular filtration rate; GLP‐1 receptor agonists, glucagon‐like peptide‐1 receptor agonists; HbA1c, hemoglobin A1c; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low‐density lipoprotein cholesterol; mRS, modified Rankin scale; SGLT2is, sodium‐glucose cotransporter 2 inhibitors; TG, triglycerides; α‐GI, alpha‐glucosidase inhibitor.

Table 4.

Logistic regression analysis of significant risk factors for severe ATBI at discharge in patients with type 2 diabetes with available laboratory data

Characteristic	Odds ratio	95% CI	P‐value
Sex, n (%)
Male	Reference	–	–
Female	1.38	1.26–1.52	<0.001
Age
/10 year increase	1.65	1.57–1.73	<0.001
BMI
≥25	Reference	–	–
<25	1.09	0.99–1.20	0.09
eGFR
≥60	Reference	–	–
31–59	0.88	0.79–0.98	0.02
<30	1.19	1.04–1.35	0.01
HbA1c
≥7.0	Reference	–	–
<7.0	0.90	0.82–0.99	0.04
HDL‐C
<40	Reference	–	–
≥40	0.63	0.54–0.72	<0.001
LDL‐C
≥120	Reference	–	–
<120	0.92	0.84–1.01	0.08
TG
≥150	Reference	–	–
<150	1.34	1.22–1.47	<0.001
Tissue plasminogen activator use	–	–	–
Anti‐diabetic agent use
DPP‐4 inhibitors	1.22	1.08–1.38	<0.001
Glinides	0.98	0.74–1.30	0.91
GLP‐1 receptor agonists	1.54	0.88–2.69	0.13
Insulin	1.86	1.66–2.09	<0.001
Metformin	0.88	0.75–1.05	0.15
SGLT2is	0.83	0.66–1.04	0.10
Sulfonylureas	1.14	0.96–1.36	0.14
Thiazolidinediones	–	–	–
α‐GIs	–	–	–
Antiplatelet/anticoagulant drug use
Aspirin	0.86	0.78–0.95	<0.001
P2Y12 inhibitors	0.90	0.81–1.00	0.04
Antihypertensive drug use
ACE inhibitors	–	–	–
Aldosterone receptor blocker use	1.12	0.83–1.50	0.46
ARBs	–	–	–
Calcium channel blocker	–	–	–
Loop diuretics	1.40	1.14–1.72	<0.001
Thiazide diuretics	–	–	–
Tolvaptan	1.28	0.71–2.31	0.42
α‐Receptor blockers	–	–	–
β‐Receptor blockers	1.08	0.91–1.27	0.38
Cholesterol‐lowering drugs use
Statins	0.86	0.77–0.96	0.01
Fibrates	0.93	0.62–1.39	0.73

The data show the results of the multivariate analysis. Thiazolidinediones, α‐GIs, ACE inhibitors, ARBs, calcium channel blockers, thiazide diuretics, and α‐receptor blockers are excluded from the multivariate analysis because they are not significant (i.e., P > 0.2) in the univariate analysis. ACE inhibitors, angiotensin‐converting enzyme inhibitors; ARBs, angiotensin II receptor blockers; BMI, body mass index; CI, confidence interval; DPP‐4 inhibitors, dipeptidyl peptidase‐4 inhibitors; eGFR, estimated glomerular filtration rate; GLP‐1 receptor agonists, glucagon‐like peptide‐1 receptor agonists; HbA1c, hemoglobin A1c; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low‐density lipoprotein cholesterol; SGLT2is, sodium glucose cotransporter 2 inhibitors; TG, triglycerides; α‐GI, alpha‐glucosidase inhibitor.

DISCUSSION

The present study shows that strong statins, fibrates, and antiplatelet agents may reduce the severity of ATBI in patients with type 2 diabetes. Many studies have suggested the pleiotropic effects of statins and fibrates; these include atheromatous plaque stabilization and anti‐inflammation, as well as lipid lowering, and these effects could be beneficial in suppressing ATBI ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ . High‐potency statins have been reported recently to reduce the carotid intima–media thickness (IMT) in various Asian countries. Fang et al. ²⁶ conducted a meta‐analysis of atorvastatin treatment for carotid IMT in Chinese patients with type 2 diabetes. They found that treatment with high‐dose atorvastatin significantly decreased the carotid IMT in patients without ischemic stroke. Yamada et al. ²⁷ reported that atorvastatin significantly reduced the relative lipid volume of carotid plaques. Our study revealed that statin treatment suppressed the severity of ATBI after its onset.

Asians are known to have a higher incidence of ischemic stroke ²⁸ , among which atherosclerosis is relatively common ²⁹ . Therefore, statin inhibition of IMT progression is directly or indirectly important in suppressing the exacerbation of atherothrombotic infarction in Asian patients. Koga et al. ³⁰ reported that long‐term administration of pravastatin could suppress IMT thickening. Strong and standard statins, when used appropriately, may be effective in inhibiting atherosclerosis. Pemafibrate use was associated with lower ORs for ATBI in the current study. Pemafibrate is a tissue‐specific PPARα agonist, known as a selective PPAR modulator alpha (SPPARMα), that significantly reduces the risk for severe ATBI. This can be attributed to two factors. SPPARMα better improves TG, HDL‐C, non‐HDL‐C, and very low‐density lipoprotein levels than do conventional fibrates ³¹ . Furthermore, pemafibrate is associated with beneficial changes in the size of atherogenic lipoproteins ³² . Patients with an mRS score of ≥3 had a slightly higher rate of concurrent statin and fibrate use (mRS score <3: n = 14, 0.03%; mRS score ≥3: n = 38, 0.09%). These patients may have experienced difficulties in lipid management, resulting in the progression of atherosclerosis and the subsequent development of severe ATBI. Aspirin and P2Y12 inhibitors, antiplatelet drugs, decreased the OR for severe ATBI. The Study of Cardiovascular Events in Diabetes trial ³³ , which was a primary prevention trial of aspirin in patients with type 2 diabetes, demonstrated a significant reduction in CVD events and a significant increase in bleeding risk.

A sub‐analysis of randomized controlled trials investigating the secondary preventive effects of P2Y12 inhibitors ³⁴ on CVD did not reach a definitive conclusion when focusing on patients with type 2 diabetes. Considering that antiplatelet drug administration for primary prevention is not recommended by the Japan Diabetes Society ¹⁹ , antiplatelet drugs cannot be recommended to prevent the exacerbation of ATBI in patients with type 2 diabetes. In this study, aldosterone receptor antagonists, loop diuretics, and tolvaptan, antihypertensive medications commonly used for their diuretic properties, were found to increase the OR for severe ATBI. Conversely, thiazide diuretics, commonly used for their antihypertensive effects, did not significantly affect the risk. Factors such as decreased extracellular fluid volume due to diuretic action and underlying conditions such as heart failure, which requires diuresis and indicates arteriosclerosis, may have influenced these results. We speculate that the significant increase in OR among insulin‐using patients is attributable to the inclusion of a large number of patients with a long history of diabetes and advanced arteriosclerosis. The utilization of DPP‐4Is was associated with a marginal, yet statistically significant, rise in the value of OR. In addition to the fact that several large clinical trials ³⁵ , ³⁶ , ³⁷ , ³⁸ did not demonstrate a superior risk of cardiovascular events with DPP‐4Is compared with placebo, the aggressive use of DPP‐4Is in the elderly and in dialysis patients, who may have more advanced atherosclerosis, may have contributed to the slight increase in the OR.

The ORs for severe ATBI increased with age in the female patients and in patients with reduced renal function. The higher OR in women than in men could be due to the following reasons. The mean age of the included patients was 74.2 years, and many women were postmenopausal. Postmenopausal women have more ATBI‐related risk factors, such as dyslipidemia, and have a declined endothelial function ³⁹ . Unfortunately, unlike men, women with type 2 diabetes tend to receive inadequate treatment, with a lower proportion achieving HbA1c control below 7% and lower rates of statin, antihypertensive, and aspirin use ⁴⁰ . This treatment bias may have contributed to a higher risk of severe ATBI in women with type 2 diabetes than in men. Elderly individuals and those with decreased renal function progress to a state of advanced arteriosclerosis and various complications ⁴¹ , ⁴² , making them more prone to ATBI exacerbation after onset. The use of tissue‐type plasminogen activator (t‐PA) significantly increases the probability of achieving an mRS score of <3. Although multiple studies have reported improvements in mRS scores with t‐PA administration, we found different results. This discrepancy may be attributed to the lower proportion of patients who received t‐PA in patients with a pre‐admission mRS score of <3 (1.3%) than in those with a pre‐admission mRS score of ≥3 (20.5%). Given that the mRS score at discharge is significantly influenced by that at pre‐admission, the bias in the patient population receiving t‐PA may have affected our findings. A similar analysis was conducted for patients with available laboratory values; however, because t‐PA was only administered to patients with an mRS score of <3 at discharge, further analysis was not feasible.

Finally, the relation of lipid‐related parameters, namely, LDL‐C, HDL‐C, and TG levels, to severe ATBI was examined. These parameters have been reported to be strongly associated with atherosclerosis. However, in the current study, LDL‐C levels, the cutoff values of which were set based on current guidelines, did not show significance, whereas TG levels <150 mg/dL were associated with an increased OR. High LDL‐C and TG levels require more aggressive statin or fibrate treatment. Moreover, LDL‐C and TG levels are lower in patients with significantly impaired renal function, suggesting the potential influence of various patient characteristics. Nevertheless, further investigation is still warranted because the reasons for this could not be clearly elucidated. In contrast, achieving HDL‐C levels within the cut‐off value significantly reduced the OR for severe ATBI. Previous studies in Japan showed an association between carotid IMT and HDL‐C levels ⁴³ . Takase et al. ⁴⁴ also reported in their community cohort study that IMT was significantly decreased with an HDL‐C level of >40 mg/dL. We speculate that achieving the cut‐off value for HDL‐C in this study might have suppressed atherosclerosis and reduced the risk of severe ATBI. Approximately 10% of the patients had available laboratory data, but the ORs for variables such as sex and age were nearly identical to those derived from the evaluation of patients without laboratory data. Consequently, it is inferred that even with an increased proportion of patients with laboratory data, the ORs in eGFR and lipid‐related parameters will exhibit similar trends.

The present study has several limitations that must be considered when interpreting the findings. First, because the MDV database primarily includes patients treated in acute care and emergency hospitals utilizing the DPC system, it may not fully represent the actual situation of all patients with type 2 diabetes. However, considering the similarity in the proportion of patients with type 2 diabetes reported in the MDV database and in the Japan Health and Nutrition Survey ⁴⁵ , the results of this study may apply to most patients with type 2 diabetes. Second, type 2 diabetes was defined based on patients diagnosed with the disease or prescribed hypoglycemic agents. By excluding patients treated solely with dietary and exercise therapy and those with early‐stage type 2 diabetes, the study likely included a higher proportion of patients with a relatively longer history of the disease. Third, the severity of ATBI was determined according to the mRS score, which may be subject to evaluator subjectivity. Therefore, prospective studies are necessary for a rigorous confirmation of the results. Finally, only 10% of the patients had available laboratory data. Given that these patients may have unknown factors that were not included in the analysis, their impact on the results cannot be ruled out. We did not conduct an analysis that included the dosage and duration of concomitant medications; therefore, the possibility that these factors might influence the results also cannot be ruled out. Furthermore, some concomitant medications may not be as frequently used, and an increase in their usage frequency could potentially affect the results, and this cannot be ruled out. Using indices such as the Diabetes Complications Severity Index (DCSI) to evaluate the severity of complications enables a more accurate assessment of their impact on severe ATBI. However, studies utilizing claims databases face challenges in accurately capturing information on all comorbid conditions, as was the case in this study. The precise calculation of the DCSI and the development of analysis methods that account for the severity of complications remain future research challenges. Patients with type 2 diabetes managed with dietary and exercise therapy could not be included. The study likely excluded patients with a short history of type 2 diabetes. Our future research direction includes prospective cohort studies based on the significant factors that contribute to the suppression of ATBI exacerbations identified in this study. In addition, we will investigate how patient education, including medication adherence and lifestyle habits, influences these outcomes.

In conclusion, actively managing lipid levels using statins and fibrates may be beneficial in preventing the exacerbation of ATBI. However, optimal medication adherence rates for dyslipidemia treatment drugs are low, possibly due to serious side effects, such as rhabdomyolysis and worsening of glycemic control in diabetes. In this study, statin and fibrate administration rates were indicating that the administration rates were not particularly high. We emphasize the need for a comprehensive multidisciplinary medical team and the provision of appropriate care for adverse events to prevent the onset and exacerbation of ATBI through the proactive use of statins and fibrates. This study identified the factors related to severe ATBI. Future research directions include using the propensity score matching method to adjust for the characteristics of patients with mRS scores of ≥3 and <3 and conducting a detailed examination.

DISCLOSURE

A.S. received lecture fees from Astellas Pharma, Inc., Eli Lilly Japan K.K., Novo Nordisk Pharma, Inc., and Sanofi K.K. The authors declare no conflicts of interest.

Approval of the research protocol: This study was conducted in accordance with the Declaration of Helsinki and Ethical Guidelines for Medical and Health Research Involving Human Subjects. The Ethics Board of Musashino University stipulated that the protocol for this study did not require ethical approval because all available data were completely anonymous with no personal information, which is characteristic of DPC‐based clinical databases.

Informed consent: All patient data were anonymized and did not contain personal data; therefore, informed consent was not required.

Approval date for registration of study/trial: N/A.

Animal studies: N/A.

Supporting information

Table S1. The frequency of use of statins and fibrates by component.

Table S2. The frequency of use of statins and fibrates by component.