A population of 5516 elderly patients with T2DM (aged ≥ 60 years) visited the Geriatric Hospital of Nanjing Medical University in Nanjing, China, between Jan 2013 and Dec 2020. The diagnosis of T2DM was based on the diagnostic criteria of the World Health Organization in 1999 (10). After excluding those who had a history of type 1 diabetes, latent autoimmune diabetes in adults (LADA), or secondary diabetes or those who had any missing data for key variables including blood pressure, HbA1c, total cholesterol (TC), low density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C) triglyceride (TG) and body mass index (BMI), a total of 3840 elderly patients with T2DM were finally eligible for analysis. Patients were divided into four groups according to diabetic duration: < 1 year (Group 1), 1~5 years (Group 2), 5~10 years (Group 3), and ≥ 10 years (Group 4). The Ethics Committee of the Geriatric Hospital of Nanjing Medical University approved the study protocol. Each patient signed written informed consent.

The clinical data of all elderly patients with T2DM were collected from the electronic medical record system. These information included diabetes duration, age, sex, education level, smoking status, family history of diabetes, systolic pressure (SBP), diastolic pressure (DBP), BMI, HbA1c, TG, TC, LDLC, HDL-C, therapeutic regimens, such as diabetes treatment (diet and exercise alone and anti-hyperglycaemic agents), and antihypertensive agents. Education level was defined as ≥ 9 years or < 9 years. Smoking status was divided into two groups, current smoking and non-current smoking. BMI was calculated as weight (kg)/height (m)². Overnight fasting venous blood samples were obtained to measure HbA1c. All blood samples were measured at the laboratory of the Geriatric Hospital of Nanjing Medical University.

The data of macrovascular and microvascular complications of diabetes were also collected based on the available information. The diagnosis of coronary heart disease (CHD) was based on a history of angina pectoris, myocardial infarction, percutaneous coronary intervention or coronary artery bypass. The diagnosis of peripheral artery disease (PAD) was based on an ankle-brachial pressure index (ABI) < 0.9. The estimated glomerular filtration rate (eGFR) was calculated by the CKD-EPI formula (11), and the diagnosis of diabetes kidney disease (DKD) was based on the KDIGO clinical practice guidelines (12, 13). Albuminuria was defined as a urine albumin/creatinine ratio (ACR) > 30 mg/gCr. Patients with eGFR < 60 ml·min^-1 ·1.73m² or albuminuria were defined as having DKD.

According to the guidelines for Chinese T2DM (2020 edition) (6), the T2DM control targets of comprehensive management include BP<130/80 mmHg, HbA1c < 7%, TC < 4. 5 mmol/L, LDL-C < 2. 6 mmol/L, HDL-C >1.0 mmol/L (males) or >1.3 mmol/L(females), TG< 1. 7 mmol/L, and BMI < 24.0 kg/m².

Our clinical characteristics were non-normally distributed, descriptive information was presented as median (interquartile range [IQR]) for continuous variables or numbers and percentages for categorical variables. For continuous variables, Kruskal-Wallis and post hoc multiple comparison Dunn’s tests were used to test for the differences (R package FSA). For categorical variables, χ2 tests and post hoc tests were carried out to examine the significant differences (R package companion). Logistic regression analysis was used to calculate the odds ratios (ORs) and 95% confidence intervals (95% CIs) for the T2DM achieved control targets or the associations of different diabetic duration groups with the risks of diabetic macrovascular and microvascular complications. The unadjusted ORs and 95% CIs were calculated in Model 1. Two other models were progressively adjusted for age, sex, education, BMI (Model 2), smoking and family history of diabetes (Model 3). Patients with a diabetes duration < 1 year were the reference group. P values for trend were calculated by modeling ordinal categories as continuous variables. A two-tailed threshold of P < 0.05 was considered statistically significant. Missing data was presented as not available and were not taken into consideration during analysis. The statistics were calculated using R statistical software version 4.1.0.

The characteristics of elderly patients with T2DM according to different diabetic duration groups are described in ###$###

Table 1. A total of 3840 elderly patients with T2DM were classified as group 1 (n = 972), group 2 (n = 896), group 3 (n = 875) and group 4 (n = 1097). Between the four groups, there were no significant differences in age, SBP, the proportion of higher education, current smoking and patients who used antihypertensive agents (all P > 0.05). In group 4, compared with groups 1 and 2, there was a significantly decreased percentage of male (P _{group 4 vs group 1} = 0.013, P _{group 4 vs group 2 =} 0.035) and lower levels of DBP (P _{group 4 vs group 1} < 0.001, P _{group 4 vs group 2} = 0.014), BMI (P _{group 4 vs group 1} <0.001, P _{group 4 vs group 2} < 0.001), TG (P _{group 4 vs group 1} < 0.001, P _{group 4 vs group 2} < 0.001) and TC (P _{group 4 vs group 1} < 0.001, P _{group 4 vs group 2} = 0.018), whereas the increased percentage of family history of diabetes (P _{group 4 vs group 1} < 0.001, P _{group 4 vs group 2} < 0.001). With prolonged diabetic status, there was a significantly decreasing trend in the proportion of patients who used diet and exercise alone to control blood glucose (28.70%, 16.96%, 10.06%, 3.92%, P < 0.001). Compared to group 1, the level of HbA1c was significantly higher in group 4 (7.40% vs 7.10%, P = 0.015), but was significantly lower in group 2 (6.70% vs 7.10%, P < 0.001) and group 3 (7.00% vs 7.10%, P = 0.002).

The control rates of T2DM targets in elderly patients with T2DM between different groups are shown in Table 2. Compared to group 1, the control rate of HbA1c was higher in group 2 (61.61% vs 45.86%, P < 0.001), whereas the rate was significantly lower in group 4 (36.55% vs 45.86%, P < 0.001). Group 4 had a significantly higher control rate of TC when compared with groups 1 (47.77% vs 41.15%, P = 0.018). The similar results were observed for the control rates of LDL-C (47.86% vs 39.92%, P = 0.002). Patients of group 4 were more likely to be higher control rate of TG and BMI when compared with other groups (all P < 0.05). Additionally, no significant differences in the control rates of blood pressure and HDL-C were found across the groups with various durations of diabetes (all P > 0.05).

As shown in Table 3, after adjusting for age, sex, education, BMI, smoking and family history of diabetes (model 3), T2DM patients in group 2 had significantly increased OR for achieving the control target for HbA1c (OR_HbA1c = 1.92, 95% CI = 1.59-2.31, P < 0.001), while elderly T2DM patients in group 4 showed significantly decreased OR for achieving the control target for HbA1c (OR_HbA1c = 0.65, 95% CI = 0.54-0.78, P < 0.001). The significantly increased ORs for achieving the control targets for TC (OR_TC = 1.36, 95% CI =1.14-1.63, P < 0.001), LDL-C (OR_LDL-C = 1.39, 95% CI =1.17-1.66, P < 0.001) and BMI (OR_BMI = 1.82, 95% CI =1.52-2.18, P < 0.001) were observed in group 4. Interestingly, we observed the gradually increased ORs for achieving the control targets for TC with the prolonged duration of diabetes (OR_{group 2} = 1.75, 95% CI =1.45-2.12; OR_{group 3} = 1.38, 95% CI =1.13-1.68; OR_{group 4} = 1.76, 95% CI =1.46-2.14; P < 0.001) in model 3.

The prevalence of diabetic macrovascular and microvascular complications among the groups are presented in Table 4. Notably, there are 2157 (56.18%) patients with missing PAD data, 1214 (31.61%) with missing albuminuria, and 114 (2.97%) with missing eGFR. Between the four groups, there was no significant difference in the percentage of patients with eGFR<60 ml·min-1 ·1.73m² (P = 0.420). Compared to group1, group 4 had the higher percentage of patients with CHD, PAD and albuminuria (all P < 0.05). As shown in Table 5, after adjusting for potential confounders, we observed elderly T2DM patients in groups 3 and 4 had significantly increased ORs for CHD (OR_{group 3 =} 1.43, 95% CI =1.07-1.89; OR_{group 4 =} 1.56, 95% CI =1.19-2.04; P _trend< 0.001) and for PAD (OR_{group 3 =} 1.82, 95% CI =1.10-3.01; OR_{group 4 =} 1.60, 95% CI =1.00-2.55; P _trend= 0.022) in model 3. Patients in group 4 had significantly increased ORs for albuminuria (OR _albuminuria = 1.96, 95% CI =1.55-2.48, P _trend < 0.001; OR _eGFR = 1.36, 95% CI =1.05-1.75, P _trend =0.030).