Study design and participants
This was a cross-sectional study. We enrolled 623 adult (age ≥ 18 years) patients with T2DM who were in hospital due to poor glycemic control from January 2019 to July 2022 at the Department of Endocrinology, Jining First People’s Hospital that is affiliated to the Jining Medical University, Jining, China. Patients with acute pancreatitis, pancreatic atrophy, and SBP < 90 mmHg or diastolic BP (DBP) < 60 mmHg were excluded. Finally, a total of 615 patients were included in the analysis. The study protocol was approved by the Medical Ethics Committee of the Jining Medical University (No. JNMC-2022-YX-019).
Clinical and laboratory measurements
All data were extracted from hospital medical records by trained staff using the quality-control Epidata version 3.1 software.
Demographics, lifestyle, history of chronic diseases, duration of diabetes, medications used were collected. Body height and weight were measured by height and weight meters (OMRON HNH-318; OMRON Corporation, Shenzhen, China). Blood pressure was measured by an electronic BP meter (OMRON HBP-1100U; OMRON Corporation, Dalian, China) in the seated position, with feet on the floor and arm supported at heart level. Laboratory examination items, including lipid profiles and hepatorenal function, were tested enzymatically by an automatic biochemistry analyzer (AU5831, Beckman Coulter, USA). HbA1c was tested by a hemoglobin analyzer (Bio-Rad D-10) using high-pressure liquid chromatography. All information above was collected at the time of admission of the patients.
A standardized steamed bread meal test (SBMT) was conducted before patient discharge. The night before the SBMT, patients stopped taking medications that would affect the trial and kept overnight fasting. The next morning, before patients took medications, venous blood samples were drawn to measure blood glucose, insulin, and C-peptide at fasting, 60 min, 120 min, and 180 min following the ingestion of 100-g flour. Insulin and C-peptide were tested by an automatic electrochemiluminescence analyzer (Cobas e801, Roche Diagnostics, Mannheim, Germany). Blood glucose was tested enzymatically by an automatic biochemistry analyzer (AU5831, Beckman Coulter, USA).
The diagnosis of T2DM was made according to the 1999 World Health Organization criteria: fasting blood glucose (FBG) ≥ 7.0 mmol/L or 2-h oral glucose tolerance test plasma glucose ≥ 11.1 mmol/L or self-reported physician-diagnosed diabetes. Hypertension was defined as SBP ≥ 140 mmHg and/or DBP ≥ 90 mmHg, self-reported physician-diagnosed hypertension, or taking antihypertensive medications even with SBP < 140 mmHg or DBP < 90 mmHg. Overweight was defined as body mass index (BMI) of ≥ 25 kg/m2 but < 30 kg/m2, and obesity is BMI ≥ 30 kg/m2, according to the American Heart Association recommendations .
BMI was calculated as weight (kg)/[height (m) × height (m)]. Insulin secretion-sensitivity index-2 (ISSI2) = (AUCinsulin/AUCglucose) × Matsuda, where AUCinsulin and AUCglucose are the areas under the insulin curve and glucose curve within 120 min of SBMT . Matsuda = 10,000/√(MG × MI) × (FG × FI), where MG is mean glucose, MI is mean insulin, FG is fasting glucose, and FI is fasting insulin. Homeostatic model assessment 2-insulin resistance (HOMA2-IR) and homeostatic model assessment 2-beta (HOMA2-B) were calculated by HOMA2 model based on blood glucose and C-peptide . The unit of blood glucose, insulin, and C-peptide were mmol/L, μIU/ml, and nmol/L, respectively. Antihypertensive treatment rate = (number of participants undergoing oral antihypertensive agent treatment/number of participants with hypertension) × 100%.
We used spline function in the RStudio version 1.4.1106. to estimate the blood glucose, insulin, and C-peptide at 30 min of SBMT. We divided participants into two groups based on the threshold of 10% of HbA1c and performed analyses in each group.
Continuous variables with normal distribution were presented as mean ± SD and compared by two independent sample t-test or one-way analysis of variance. Continuous variables with skewed distribution were presented as median (25th–75th) and compared by Wilcoxon rank sum test or Kruskal–Wallis H test. Categorical variables were presented as numbers (percentages) and compared by Chi-square test.
To identify the association between BP and beta-cell function (ISSI2), multivariable linear regression analysis was performed with adjustment of age, sex, duration of diabetes, BMI, triglyceride (TG), low-density lipoprotein cholesterol (LDL-C), Matsuda index, and medications used. BP was analyzed in the model as a continuous variable and a categorical variable, respectively. Among participants with HbA1c ≥ 10%, we performed sensitivity analyses in following subjects: insulin sensitivity in 10th–90th percentiles, without taking oral antihypertensive agents, without accepting insulin treatment, and duration of diabetes < 1 year or ≥ 1 year.
To identify the dose–response relationship between SBP and ISSI2, we conducted restricted cubic splines (RCS) analyses. We set four knots at the 5th, 25th, 75th, and 95th percentiles, and set SBP of 120 mmHg as a reference. In addition, to determine the dose–response relationship between SBP and blood glucose, we further analyzed the association between SBP and blood glucose by RCS with the same parameters set. Poverall < 0.05 and Pnon-linear < 0.05 indicated that the association was nonlinear.
To determine whether beta-cell function mediated the association between SBP and 2 h postprandial blood glucose (PBG) in participants with HbA1c ≥ 10%, we performed mediation analyses. Bootstrap method was used to estimate the confidence interval with the seed set to 5000.
All statistical analyses were conducted in SAS version 9.4 (SAS Institute Inc, Cary, NC). A two-tailed P value < 0.05 was considered statistically significant.