Association of metformin exposure with low risks of frailty and adverse outcomes in patients with diabetes.

Background

Diabetes is an age-related chronic metabolic disease. In 2021, 537 million (10.5%) adults aged 20–79 years were estimated to live with diabetes. The number is predicted to rise to 643 million by 2030 and 783 million by 2045 [ 1 ]. The overall prevalence of diabetes diagnosed by World Health Organization criteria in Chinese adults was 11.2% in 2017, which was increased by 1.5% over the past 10 years [ 2 ]. Diabetes is associated with many complications, such as macrovascular and microvascular target organ damage, leading to serious public health problems. Frailty is a clinical state with multiple decreased functions and more vulnerability to stress, resulting in a high risk of adverse outcomes, such as falls, disability, hospitalization, dependency, and even death [ 3 , 4 ]. Diabetes, a risk factor for frailty, can increase the prevalence of frailty [ 3 , 5 ]. Meanwhile, frailty also worsens health conditions in older adults with diabetes [ 6 ]. Research has found that patients with diabetes are more likely to exhibit reduced leg muscle mass with function loss [ 7 , 8 ], which might be associated with frailty performance. Chronic inflammation and insulin resistance are two of the most common pathological mechanisms of diabetes and frailty [ 8 ].

Metformin is a strongly recommended first-line drug for the treatment of diabetes as well as a potential anti-aging medicine. Evidence from both clinical and pre-clinical studies supports the benefits of metformin in improving healthspan and expanding lifespan [ 9 ]. Moreover, metformin can reduce the risk of all-cause mortality and age-related diseases, such as cardiovascular disease and cancer in non-diabetic adults [ 10 ]. These anti-aging effects are mainly from non-diabetes research, which is associated with improvement in deregulated nutrient sensing, altered intercellular communication, proteostasis disorders, and genomic instability [ 11 ]. In addition, metformin inhibits the hallmarks of aging such as mitochondrial dysfunction, cellular senescence, and stem cell exhaustion [ 11 ].

Diabetes and frailty often co-exist in adults. Only a few studies have investigated the relationship between frailty and metformin use, and have found that metformin was associated with a low risk of frailty evaluated by the Frailty Index-40 and frailty-related diseases in older male veterans with diabetes [ 12 , 13 ]. However, the protective effects of metformin on frailty-related skeletal muscle remain controversial [ 14 – 16 ]. Although previous studies have shown that metformin improves multiple aging phenotypes, whether it has a similar effect on age-related frailty requires further investigation. Our objective was mainly to explore the association between metformin use and frailty in patients with diabetes.

Methods

Study design and participants

Data were obtained from the ongoing cohort of the Physiological Model for Frailty and Resilience Study (SMART). The inclusion criteria were a clinical diagnosis of type 2 diabetes and patients aged 40 years or older. Patients with type 1 diabetes or severe multiple organ dysfunction were excluded. The clinical data for the first participant were collected in January 2018. Ultimately, 422 participants met the inclusion criteria. The median follow-up time was 21.0 months (95% CI 19.3–22.7) for participants in the study. All participants voluntarily participated in this study and signed an informed consent form. The study design was approved by the Ethics Review Board of Xuanwu Hospital Capital Medical University. Information was only collected after written consent was obtained from all participants. This study was conducted in accordance with the Declaration of Helsinki.

Metformin exposure

Individuals treated with metformin were considered exposed, including monotherapy and a combination of other antidiabetic therapy. Those using sulfonylureas, glinides, glucosidase inhibitors, thiazolidinediones, dipeptidyl peptidase 4 (DPP-4) inhibitors, sodium–glucose transport protein 2 (SGLT2) inhibitors, glucagon-like peptide-1 (GLP-1) agonist, insulin, and other antidiabetics, or lifestyle management (exercise and/or diet control) were considered as not exposed.

Measurements

General characteristics of the participants, including age and sex, duration of diabetes, comorbidities, and/or target organ damage, were obtained from electronic medical records. Weight, height, waist, and hip were measured on admission, and the body mass index (BMI) (kg/m 2 ) and waist-to-hip ratio were calculated.

Frailty assessment

Frailty was assessed by Fried phenotype, which consists of five items [ 3 ]. Weakness was evaluated by grip strength with a lowest 20% of the Chinese older population adjusted by sex and BMI [ 17 ]. It was measured with a handgrip dynamometer (CAMRY, EH101) in this study. Slowing walking speed was evaluated by the 4 m walking test with a lowest 20% of the Chinese older adults adjusted by sex and height [ 18 ]. Exhaustion was identified by self-report using two questions from the Centre of Epidemiology Studies Depression Scale [ 19 ]. Inactivity was defined as little or no activity, or less than 3 h of outdoor activity per week. Weight loss was measured by unintentionally losing ≥ 4.5 kg or 5.0% in the past year or BMI < 18.5 kg/m 2 . Frailty was defined as the presence of three or more of the above phenotypes, and non-frailty was that met 0 to 2 items.

Quality of life assessment

Medical Outcomes Study’s 36-Item Short Form Health Survey (SF-36) is a widely used method to evaluate the quality of life and is also suitable for Chinese adults [ 20 ]. It includes eight dimensions of physical functioning, role physical, body pain, general health, vitality, social functioning, role-emotional, and mental health. Health transition represents the overall health changes in the past year. The SF-36 total scores are the sum of each dimension [ 21 , 22 ]. The higher the SF-36 score, the worse quality of life.

Laboratory tests

Fasting blood and morning urine were collected. The levels of fasting plasma glucose, hemoglobin A1C, triglyceride, total cholesterol, high-density lipoprotein, low-density lipoprotein, creatinine, blood urea nitrogen, prealbumin, albumin, hemoglobin, C-reactive protein (CRP), D-dimer, fibrinogen, and N-terminal pro-B-type natriuretic peptide (NT-proBNP) in serum were assayed. Urine microalbuminuria/creatinine (mALB/Cre) was calculated.

Outcomes

The primary outcome was the combined endpoint of cardiovascular events, cerebrovascular events, readmission, and death. The cardiovascular events were assessed by Major Adverse Cardiovascular Events (MACE), which include unstable angina, acute myocardial infarction, severe arrhythmia, heart failure, and coronary heart disease death in this study. The cerebrovascular events included acute stroke and stroke death. For those patients already with a history of coronary artery disease and stroke at baseline, new-onset acute cardiovascular and cerebrovascular events during the follow-up period are used as the criterion. Readmission was defined as hospitalization within 30 days and hospitalization due to acute cardiovascular and cerebrovascular events during follow-up. All information was collected from patients and medical system records.

Statistical analysis

Continuous variables were described with a mean (± standard deviation) or median (quartile) and analyzed by Student’s t-test and Mann–Whitney U test. Frequency (%) and Chi-square test were used to describe and analyze the categorical variables, respectively. Binary logistic regression was used to analyze the independent effects of metformin on frailty and reported the odds ratios (OR) and 95% confidence intervals (CIs). Furthermore, survival analysis was estimated by the Kaplan–Meier method and the log-rank test was used to evaluate the difference between the metformin and non-metformin groups. Cox proportional-hazards models were used to assess the effects of metformin on the survival of patients with diabetes by adjusting age, sex, and frailty. Baseline variables that were considered clinically relevant or that showed a univariate relationship with outcome were entered into logistic regression analysis and multivariate Cox proportional-hazards regression model. Data were analyzed using IBM SPSS version 26.0 software (Inc., Chicago, IL, USA) and GraphPad Prism 7.0 software (GraphPad Software Inc., CA, USA). P -value < 0.05 is defined as a statistically significant difference.

Results

General characteristics in non-frail and frail participants with diabetes

The mean age was 70.35 (± 10.71) years, and 64.0% ( n = 270) were men. The prevalence of frailty was 19.4% ( n = 82) in participants with diabetes. The general characteristics of the non-frail and frail groups are compared in Table 1 . Compared with non-frail patients, frail patients were older and had a higher prevalence of hypertension, coronary atherosclerotic heart disease, chronic kidney disease, stroke, osteoarthritis, cancer, peripheral vascular disease, diabetic peripheral neuropathy, and polypharmacy. Frail patients had lower grip strength, slower walking speed, and worse SF-36 scores than that non-frail. A total of 245 patients with diabetes were treated with metformin, with 32 (39.0%) in the frail group and 213 (62.6%) in the non-frail group (Table 1 ). However, there was no difference in other antidiabetic treatments and duration of diabetes between the two groups. Table 1 Characteristics of non-frail and frail participants with diabetes Variable Non-frail ( n = 340) Frail ( n = 82) P value Age (years) 67.84 ± 9.60 80.76 ± 8.65 < 0.001 Female (n, %) 124 (36.5) 28 (34.1) 0.694 Smoking (n, %) 158 (46.5) 36 (43.9) 0.675 Drinking (n, %) 130 (38.3) 23 (28.4) 0.094 WHR 0.95 ± 0.09 1.08 ± 1.03 0.884 BMI (kg/m 2 ) 25.85 ± 3.19 25.25 ± 4.10 0.226 Duration (years) 11.18 ± 8.62 12.43 ± 8.67 0.265 Antidiabetic treatments Lifestyle management (n, %) 42 (12.4) 14 (17.1) 0.258 Metformin (n, %) 213 (62.6) 32 (39.0) < 0.001 Sulfonylureas (n, %) 60 (17.6) 14 (17.1) 0.902 Glinide (n, %) 17 (5.0) 6 (7.3) 0.576 Glucosidase inhibitor (n, %) 140 (41.2) 40 (48.8) 0.211 Thiazolidinedione (n, %) 21 (6.2) 8 (5.6) 0.250 DPP-4 inhibitors (n, %) 27 (7.9) 3 (3.7) 0.176 SGLT2 inhibitors (n, %) 19 (5.6) 4 (4.9) 0.799 GLP-1 agonist (n, %) 7 (2.1) 0 (0.0) 0.407 Insulin (n, %) 100 (29.4) 21 (25.6) 0.494 Other medicines (n, %) 3 (0.9) 1 (1.2) 0.580 Comorbidity Hypertension (n, %) 256 (75.3) 74 (90.2) 0.003 CAD (n, %) 111 (32.6) 50 (61.0) < 0.001 COPD (n, %) 13 (3.8) 5 (6.1) 0.542 CKD (n, %) 34 (10.0) 22 (26.8) < 0.001 Stoke (n, %) 58 (17.1) 30 (36.6) < 0.001 Osteoarthritis (n, %) 52 (15.3) 29 (35.4) < 0.001 Cancer (n, %) 32 (9.4) 23 (28.0) < 0.001 Target organ damage PVD (n, %) 43 (12.6) 23 (28.0) 0.001 DR (n, %) 45 (13.2) 7 (8.5) 0.245 DPN (n, %) 44 (12.9) 19 (23.2) 0.020 Polypharmacy (n, %) 203 (59.7) 68 (82.9) < 0.001 Grip strength (kg) 30.44 ± 9.52 18.24 ± 7.07 < 0.001 Walking speed (m/s) 1.01 ± 0.26 0.59 ± 0.22 < 0.001 SF-36 score 100.65 ± 11.36 115.49 ± 8.53 < 0.001 Data for continuous variables are presented as mean ± (standard deviation) or median (interquartile range). Data for categorical variables are presented as n (percentage) WHR, waist-to-hip ratio; BMI, body mass index; DPP-4, dipeptidyl peptidase 4; SGLT2, sodium–glucose transport protein 2; GLP-1, glucagon-like peptide-1; CAD, coronary atherosclerotic heart disease; COPD, chronic obstructive pulmonary disease; CKD, chronic kidney disease; PVD, peripheral vascular disease; DR, diabetic retinopathy; DPN, diabetic peripheral neuropathy

Table 2 exhibits the laboratory test results of the non-frail and frail groups. Compared with non-frail participants, frail patients had higher levels of creatinine, blood urea nitrogen, urine mALB/Cre and lower levels of albumin and hemoglobin. Higher D-dimer, fibrinogen, and NT-proBNP levels were observed in diabetic patients with frailty. However, no difference was found between the two groups regarding blood glucose and lipid biomarkers. In accordance with these variables, univariate binary logistic regression analysis of risk factors for frailty showed similar findings (Additional file 1 : Table S1). Table 2 Laboratory tests of non-frail and frail patients with diabetes Variable Non-frail ( n = 340) Frail ( n = 82) P value FPG (mmol/L) 7.18 ± 2.36 7.16 ± 3.27 0.079 HbA1c (%) 7.37 ± 1.52 7.06 ± 1.33 0.067 Triglyceride (mmol/L) 1.79 ± 0.79 1.47 ± 0.70 0.379 Total cholesterol (mmol/L) 4.25 ± 1.15 4.04 ± 1.21 0.044 HDL (mmol/L) 1.12 ± 0.32 1.07 ± 0.30 0.214 LDL (mmol/L) 2.48 ± 0.90 2.39 ± 1.00 0.242 Creatinine (umol/L) 70.80 ± 40.76 94.89 ± 64.03 < 0.001 BUN (mmol/L) 6.19 ± 2.19 8.92 ± 4.96 < 0.001 Prealbumin (mg/L) 248.31 ± 121.64 222.26 ± 65.25 0.062 Albumin (mg/L) 40.14 ± 3.88 37.29 ± 5.69 < 0.001 Hemoglobin (g/L) 135.98 ± 17.41 124.27 ± 21.90 < 0.001 CRP (mg/L) 2.13 (1.90, 4.00) 3.00 (2.00, 5.00) 0.222 D-Dimer (ug/mL) 0.29 (0.20, 0.45) 0.59 (0.37, 1.00) < 0.001 Fibrinogen (g/L) 3.38 ± 0.72 3.66 ± 0.98 0.016 NT-proBNP (pg/mL) 74.85 (39.75, 197.25) 348.5 (103.98, 781.75) < 0.001 Urine mALB/Cre 1.50 (0.30, 7.30) 7.00 (0.91, 28.50) < 0.001 Data for continuous variables are presented as mean ± (standard deviation) or median (interquartile range) FPG, fasting plasma glucose; HbA1c, hemoglobin A1c; HDL, high-density lipoprotein; LDL, low-density lipoprotein; BUN, blood urea nitrogen; CRP, C-reactive protein; NT-proBNP, N-terminal pro-B-type natriuretic peptide; mALB/Cre, microalbuminuria/creatinine

Metformin exposure in different frailty groups

Compared with the non-metformin group, the metformin group was younger and had a higher BMI, a higher risk of coronary atherosclerotic heart disease, chronic kidney disease, osteoarthritis, and peripheral vascular disease, a lower proportion of frailty, greater grip strength, and faster walking speed (Table 3 ). However, these differences in physical performance did not exist in the subgroup analysis of the non-frail and frail participants (Table 4 ). Both frail and non-frail patients treated with metformin had a lower prevalence of coronary atherosclerotic heart disease and chronic kidney disease than those not treated with metformin (Tables 3 and 4 ). Non-frail participants taking metformin had a lower prevalence of osteoarthritis and peripheral vascular disease and higher levels of fasting plasma glucose, hemoglobin A1c, triglyceride, prealbumin, albumin, and hemoglobin than those not taking metformin (Tables 4 and 5 ), which was in accordance with the results in Table 6 . However, these differences did not exist in the frail group. Frail patients treated with metformin had lower levels of CRP than that with non-metformin (Table 5 ). Participants taking metformin were more likely to have lower D-dimer and NT-proBNP levels (Table 5 ) and better quality of life assessed by the SF-36 score (Table 4 ), consistent with the findings in Table 6 and 3 , respectively. Table 3 Characteristics of non-metformin and metformin groups Variable Non-metformin ( n = 177) Metformin ( n = 245) P value Age (years) 74.80 ± 11.01 67.14 ± 9.27 < 0.001 Female (n, %) 62 (35.0) 90 (36.7) 0.719 Smoking (n, %) 76 (42.9) 118 (48.2) 0.288 Drinking (n, %) 53 (29.9) 100 (41.2) 0.018 WHR 1.00 ± 0.65 0.95 ± 0.10 0.305 BMI (kg/m 2 ) 25.07 ± 3.64 26.21 ± 3.10 0.001 Duration (years) 11.12 ± 9.20 11.65 ± 8.20 0.538 Antidiabetic treatments Sulfonylureas (n, %) 29 (16.4) 45 (18.4) 0.597 Glinide (n, %) 11 (6.2) 12 (4.9) 0.557 Glucosidase inhibitor (n, %) 78 (44.1) 102 (41.6) 0.618 Thiazolidinedione (n, %) 7 (4.0) 22 (9.0) 0.044 DPP-4 inhibitors (n, %) 9 (5.1) 21 (8.6) 0.169 SGLT2 inhibitors (n, %) 3 (1.7) 20 (8.2) 0.004 GLP-1 agonist (n, %) 2 (1.1) 5 (2.0) 0.704 Insulin (n, %) 51 (28.8) 70 (28.6) 0.957 Other medicines (n, %) 1 (0.6) 3 (1.2) 0.643 Comorbidity Hypertension (n, %) 145 (81.9) 185 (75.5) 0.116 CAD (n, %) 88 (49.7) 73 (29.8) < 0.001 COPD (n, %) 11 (6.2) 7 (2.9) 0.092 CKD (n, %) 40 (22.6) 16 (6.5) < 0.001 Stoke (n, %) 39 (22.0) 49 (20.0) 0.612 Osteoarthritis (n, %) 45 (25.4) 36 (14.7) 0.006 Cancer (n, %) 27 (15.3) 28 (11.4) 0.249 Target organ damage PVD (n, %) 38 (21.5) 28 (11.4) 0.005 DR (n, %) 17 (9.6) 35 (14.3) 0.149 DPN (n, %) 22 (12.4) 41 (16.7) 0.221 Polypharmacy (n, %) 118 (66.7) 153 (62.4) 0.372 Frailty (n, %) 50 (28.2) 32 (13.1) < 0.001 Grip strength (kg) 26.70 ± 10.43 29.30 ± 10.02 0.014 Walking speed (m/s) 0.91 ± 0.30 0.98 ± 0.29 0.038 Exhaustion (n, %) 77 (46.7) 81 (34.9) 0.018 Inactivity (n, %) 56 (35.2) 42 (19.7) < 0.001 Weight Loss (n, %) 25 (14.8) 30 (12.7) 0.547 SF-36 score 109.22 ± 12.02 115.16 ± 9.11 < 0.001 Data for continuous variables are presented as mean ± (standard deviation) or median (interquartile range). Data for categorical variables are presented as n (percentage) WHR, waist-to-hip ratio; BMI, body mass index; DPP-4, dipeptidyl peptidase 4; SGLT2, sodium–glucose transport protein 2; GLP-1, glucagon-like peptide-1; CAD, coronary atherosclerotic heart disease; COPD, chronic obstructive pulmonary disease; CKD, chronic kidney disease; PVD, peripheral vascular disease; DR, diabetic retinopathy; DPN, diabetic peripheral neuropathy Table 4 Subgroup analysis of characteristics of non-frail and frail patients with diabetes Variable Non-frail ( n = 340) Frail ( n = 82) Non-metformin ( n = 127) Metformin ( n = 213) P value Non-metformin ( n = 50) Metformin ( n = 32) P value Age (years) 71.2 ± 10.39 65.84 ± 8.51 < 0.001 83.95 ± 6.22 75.78 ± 9.62 < 0.001 Female (n, %) 44 (34.6) 80 (37.6) 0.589 18 (36.0) 10 (31.3) 0.658 WHR 0.95 ± 0.07 0.95 ± 0.10 0.220 1.14 ± 1.28 0.97 ± 0.06 0.211 BMI (kg/m 2 ) 25.07 ± 3.44 26.32 ± 2.94 0.001 25.05 ± 4.17 25.55 ± 4.03 0.582 Smoking (n, %) 55 (43.3) 103 (48.4) 0.366 21 (42.0) 15 (46.9) 0.664 Drinking (n, %) 45 (35.4) 85 (40.1) 0.393 8 (16.0) 15 (48.4) 0.002 Duration (years) 10.44 ± 9.34 11.63 ± 8.12 0.048 12.85 ± 8.60 11.78 ± 8.87 0.415 Comorbidity Hypertension (n, %) 100 (78.7) 156 (73.2) 0.255 45 (90.0) 29 (90.6) 0.926 CAD (n, %) 53 (41.7) 58 (27.2) 0.006 35 (70.0) 15 (46.9) 0.036 COPD (n, %) 6 (4.7) 7 (3.3) 0.706 5 (100.0) 0 (0.0) 0.078 CKD (n, %) 20 (15.7) 14 (6.6) 0.006 20 (40.0) 2 (6.3) 0.001 Stoke (n, %) 24 (18.9) 34 (16.0) 0.486 15 (30.0) 15 (46.9) 0.122 Osteoarthritis (n, %) 26 (20.5) 26 (12.2) 0.041 19 (38.0) 10 (31.3) 0.533 Cancer (n, %) 10 (7.9) 22 (10.3) 0.453 17 (34.0) 6 (18.8) 0.134 Target organ damage PVD (n, %) 22 (17.3) 21 (9.9) 0.045 16 (32.0) 7 (21.9) 0.319 DR (n, %) 13 (10.2) 32 (15.0) 0.208 4 (8.0) 3 (9.4) 0.564 DPN (n, %) 13 (10.2) 31 (14.6) 0.251 9 (18.0) 10 (31.3) 0.165 Polypharmacy (n, %) 77 (60.6) 126 (59.2) 0.789 41 (82.0) 27 (84.4) 0.780 Grip strength (kg) 29.86 ± 9.53 30.78 ± 9.52 0.383 17.62 ± 7.08 19.09 ± 7.10 0.304 Walking speed (m/s) 0.99 ± 0.26 1.02 ± 0.27 0.361 0.55 ± 0.19 0.64 ± 0.24 0.109 SF-36 score 113.32 ± 10.14 116.76 ± 7.16 0.008 98.10 ± 9.46 104.34 ± 12.94 0.026 Data for continuous variables are presented as mean ± (standard deviation) or median (interquartile range). Data for categorical variables are presented as n (percentage) WHR, waist-to-hip ratio; BMI, body mass index; CAD, coronary atherosclerotic heart disease; COPD, chronic obstructive pulmonary disease; CKD, chronic kidney disease; PVD, peripheral vascular disease; DR, diabetic retinopathy; DPN, diabetic peripheral neuropathy Table 5 Subgroup analysis of laboratory tests of frail and non-frail patients with diabetes Variable Non-frail ( n = 340) Frail ( n = 82) Non-metformin ( n = 127) Metformin ( n = 213) P value Non-metformin ( n = 50) Metformin ( n = 32) P value FPG (mmol/L) 6.44 ± 2.02 7.63 ± 2.44 < 0.001 7.46 ± 3.58 6.69 ± 2.70 0.497 HbA1c (%) 6.97 ± 1.32 7.61 ± 1.58 < 0.001 6.82 ± 0.96 7.45 ± 1.70 0.244 Triglyceride (mmol/L) 1.52 ± 1.03 1.96 ± 2.11 0.009 1.42 ± 0.63 1.56 ± 0.80 0.509 Total cholesterol (mmol/L) 4.20 ± 1.16 4.28 ± 1.15 0.478 4.05 ± 0.17 4.02 ± 1.31 0.686 HDL (mmol/L) 1.18 ± 0.38 1.09 ± 0.26 0.060 1.05 ± 0.31 1.10 ± 0.27 0.354 LDL (mmol/L) 2.48 ± 0.95 2.48 ± 0.87 0.747 2.39 ± 0.94 2.39 ± 1.11 0.853 Creatinine (umol/L) 78.43 ± 61.16 66.29 ± 19.43 0.084 102.46 ± 74.29 83.06 ± 41.84 0.172 BUN (mmol/L) 6.46 ± 2.70 6.04 ± 1.80 0.120 9.84 ± 5.86 7.49 ± 2.56 0.101 Prealbumin (mg/L) 228.33 ± 49.16 260.22 ± 147.79 < 0.001 219.1 ± 74.34 227.19 ± 48.47 0.704 Albumin (mg/L) 39.67 ± 4.15 40.41 ± 3.68 0.033 36.94 ± 5.96 37.84 ± 5.28 0.456 Hemoglobin (g/L) 134.06 ± 22.03 137.13 ± 13.90 0.016 123.16 ± 18.36 126.00 ± 26.74 0.287 CRP (mg/L) 2.00 (1.94, 4.00) 2.19 (1.85, 4.00) 0.822 3.47 (2.00, 5.19) 2.00 (1.64, 3.07) 0.017 D-Dimer (ug/mL) 0.33 (0.25, 0.54) 0.26 (0.19, 040) < 0.001 0.73 (0.48, 1.53) 0.41 (0.32, 0.70) 0.002 Fibrinogen (g/L) 3.41 ± 0.74 3.36 ± 0.71 0.620 3.71 ± 0.71 3.60 ± 0.84 0.945 NT-proBNP (pg/mL) 128.00 (50.03, 241.50) 57.90 (35.98, 127.25) < 0.001 493.50 (221.25, 1367.50) 178.00 (58.60, 394.50) 0.004 Urine mALB/Cre 1.43 (0.35, 8.36) 1.60 (0.30, 7.30) 0.770 3.80 (0.68, 52.80) 17.40 (1.60, 27.10) 0.723 Data for continuous variables are presented as mean ± (standard deviation) or median (interquartile range) FPG, fasting plasma glucose; HbA1c, hemoglobin A1c; HDL, high-density lipoprotein; LDL, low-density lipoprotein; BUN, blood urea nitrogen; CRP, C-reactive protein; NT-proBNP, N-terminal pro-B-type natriuretic peptide; mALB/Cre, microalbuminuria/creatinine Table 6 Laboratory tests of non-metformin and metformin groups Variable Non-metformin ( n = 177) Metformin ( n = 245) P value FPG (mmol/L) 6.73 ± 2.59 7.50 ± 2.49 0.002 HbA1c (%) 6.92 ± 1.23 7.59 ± 1.59 < 0.001 Triglyceride (mmol/L) 1.49 ± 0.93 1.90 ± 1.99 0.005 Total cholesterol (mmol/L) 4.16 ± 1.16 4.25 ± 1.17 0.442 HDL (mmol/L) 1.14 ± 0.37 1.09 ± 0.26 0.086 LDL (mmol/L) 2.45 ± 0.94 2.47 ± 0.90 0.819 Creatinine (umol/L) 85.22 ± 65.82 68.45 ± 24.14 0.001 BUN (mmol/L) 7.41 ± 4.14 6.23 ± 1.98 < 0.001 Prealbumin (mg/L) 225.72 ± 57.32 255.91 ± 139.28 0.007 Albumin (mg/L) 38.90 ± 4.87 40.08 ± 4.01 0.007 Hemoglobin (g/L) 130.98 ± 21.57 135.67 ± 16.51 0.016 CRP (mg/L) 2.34 (2.00, 4.58) 2.05 (1.84, 4.00) 0.823 D-Dimer (ug/mL) 0.41 (0.26, 0.70) 0.28 (0.19, 0.43) 0.001 Fibrinogen (g/L) 3.50 ± 0.85 3.39 ± 0.73 0.158 NT-proBNP (pg/mL) 188.00 (73.5, 495.25) 67.05 (37.00, 181.50) 0.007 Urine mALB/Cre 1.64 (0.44, 14.80) 1.85 (0.30, 10.03) 0.021 Data for continuous variables are presented as mean ± (standard deviation) or median (interquartile range) FPG, fasting plasma glucose; HbA1c, hemoglobin A1c; HDL, high-density lipoprotein; LDL, low-density lipoprotein; BUN, blood urea nitrogen; CRP, C-reactive protein; NT-proBNP, N-terminal pro-B-type natriuretic peptide; mALB/Cre, microalbuminuria/creatinine

Association between metformin and frailty

Frailty was considered the dependent variable in the study, and regression models were constructed to explore the relationship between metformin and frailty. Univariate regression analysis found that metformin was negatively associated with frailty (B = − 0.963, OR = 0.382; 95%CI 0.233–0.626) (Additional file 1 : Table S1). After adjusting for age, sex, diabetes duration, fasting plasma glucose, hemoglobin A1c, peripheral vascular disease, diabetic retinopathy, diabetic peripheral neuropathy, hypertension, coronary atherosclerotic heart disease, chronic obstructive pulmonary disease, chronic kidney disease, stroke, osteoarthritis, cancer, and polypharmacy, metformin exposure remains negatively associated with frailty (B = − 0.572, OR = 0.564; 95% CI 0.321–0.991), independently other risk factors for frailty (Table 7 ). Table 7 Regression models of the association of metformin and frailty Variable OR 95% CI P value Age 4.295 0.970–19.009 0.055 CAD 1.926 1.111–3.339 0.020 CKD 2.129 1.062–4.265 0.033 Stroke 2.875 1.560–5.298 0.001 Osteoarthritis 2.032 1.111–3.719 0.021 Cancer 3.345 1.686–6.638 0.001 PVD 2.148 1.111–4.154 0.023 Metformin 0.564 0.321–0.991 0.047 Constant 0.024 0.000 Adjusted by age, sex, duration, FPG, HbA1c, PVD, DR, DPN, hypertension, CAD, COPD, CKD, stroke, osteoarthritis, cancer, and polypharmacy. Sex, hypertension, COPD, duration, FPG, HbA1c, DR, DPN, and polypharmacy were not included in the equation FPG, fasting plasma glucose; HbA1c, hemoglobin A1c; BMI, body mass index; PVD, peripheral vascular disease; DR, diabetic retinopathy; DPN, diabetic peripheral neuropathy; CAD, coronary atherosclerotic heart disease; COPD, chronic obstructive pulmonary disease; CKD, chronic kidney disease

Metformin and frailty on adverse outcomes

The follow-up findings indicated that frail patients had higher rates of cardiovascular events (20.7% vs. 6.8%, p < 0.001), readmission (47.6% vs. 27.6%, p < 0.001), mortality (7.3% vs. 0.6%, p < 0.001), and combined primary outcomes (58.5% vs. 31.8%, p < 0.001) than non-frail patients. Subgroup analysis showed that metformin use was associated with a lower prevalence of cardiovascular events (1.9% vs. 15.1%, p < 0.001) and combined primary outcomes (26.8% vs. 40.2%, p = 0.010) in non-frail diabetic patients, whereas metformin use was only associated with a low prevalence of cardiovascular events (3.1% vs. 32.0%, p = 0.002) in frail patients with diabetes.

Kaplan–Meier analysis further indicated that the median survival time with and without metformin was 46.0 months (95% CI 30.52–61.48) and 24.0 months (95% CI 15.08–32.92), respectively. A significant difference in survival time was found between the non-frail and frail groups in patients with diabetes (χ2 = 30.127, p < 0.001; Fig. 1 A). Log-rank tests indicated a statistically significant difference in survival time between the metformin and non-metformin treatment groups (χ2 = 14.266, p < 0.001; Fig. 1 B). A similar significant difference also existed in non-frail patients with diabetes (χ2 = 6.492, p = 0.011; Fig. 1 C) but not in frail participants (χ2 = 1.653, p = 0.199; Fig. 1 D). In addition, Cox proportional hazard models showed that metformin treatment was independently and inversely associated with diabetes. Compared with non-metformin treatment, metformin use had a lower risk of combined primary outcome after adjusting for age and sex (hazard ratio (HR) = 0.558; 95% CI 0.407–0.765). When adjusted for frailty, duration, fasting plasma glucose and hemoglobin A1c, target organ damage, comorbidities, and polypharmacy, the protective effects remained (HR = 0.695; 95%CI 0.501–0.964). In addition, the results also indicated that frailty (HR = 2.009; 95%CI 1.408–2.865) and coronary atherosclerotic heart disease (HR = 1.699, 95%CI 1.226–2.356) were associated with the increased risk of combined primary outcomes (Table 8 ). In the non-frail group, metformin use was also associated with a lower risk of combined primary outcomes adjusted for age and sex (HR = 0.620; 95%CI 0.425–0.905). However, the protective effect of metformin on adverse outcomes was not found in frail participants with diabetes. Fig. 1 Kaplan–Meier survival curves for the combined primary outcome in months. A Survival time according to non-frail ( n = 340) and frail ( n = 82) diabetic patients in the total population. B Survival time according to treated with non-metformin ( n = 177) and metformin ( n = 245) in total population. C Survival time according to treated with non-metformin ( n = 127) and metformin ( n = 213) in non-frail diabetic patients. D Survival time according to treated with non-metformin ( n = 50) and metformin ( n = 32) in frail diabetic patients. P values by log-rank test Table 8 Cox regression models of metformin on the primary outcomes Variable OR 95% CI P value Frailty 2.009 1.408–2.865 < 0.001 CAD 1.699 1.226–2.356 0.001 Metformin 0.695 0.501–0.964 0.029 Adjusted for age, sex, frailty, duration, FPG, HbA1c, PVD, DR, DPN, hypertension, CAD, COPD, CKD, stroke, osteoarthritis, cancer, and polypharmacy. Age, sex, duration, FPG, HbA1c, PVD, DR, DPN, hypertension, COPD, CKD, stroke, osteoarthritis, cancer, and polypharmacy were not included in the equation FPG, fasting plasma glucose; HbA1c, hemoglobin A1c; PVD, peripheral vascular disease; DR, diabetic retinopathy; DPN, diabetic peripheral neuropathy; CAD, coronary atherosclerotic heart disease; COPD, chronic obstructive pulmonary disease; CKD, chronic kidney disease

Discussion

The frailty prevalence among diabetic patients in this study was 19.4%, which was close to the 19.32% for diabetic patients in the Beijing longitudinal study of aging II [ 5 ]. Combined coronary atherosclerotic heart disease, chronic kidney disease, stroke, osteoarthritis, cancer, and peripheral vascular disease were independent risk factors for frailty in diabetic patients. Studies on older male veterans with diabetes have indicated that metformin promoted a healthy lifespan by preventing frailty and reducing the risk of age-related comorbidities and frailty-related diseases [ 13 , 23 ]. In the present study, a total of 245 (58.1%) patients with diabetes were treated with metformin, 64 were on monotherapy and the other 181 participants were treated with combinations of metformin and other antidiabetic drugs. Among 177 patients not treated with metformin, 49 (27.7%) of them were treated with lifestyle interventions, 9 (5.1%) of them discontinued metformin due to gastrointestinal intolerance, and 119 (67.2%) were treated with other antidiabetic drugs because of disease conditions and metformin contraindications. The prevalence of gastrointestinal intolerance in our study was similar to the results of Genetics of Diabetes Audit and Research Tayside Study [ 24 ]. Consistent with previous findings [ 23 , 25 ], we also found metformin independently and negatively associated with frailty evaluated by the Fried phenotype.

In addition, metformin use could protect against age-related musculoskeletal disorders. A study suggested that metformin treatment improved the mean walking speed in pre-frail older patients without diabetes. But they did not find any influence of metformin use on grip strength, quality of life, and myostatin serum level [ 14 ]. Other studies argued that metformin weakened the benefits of physical activity, such as resistance, endurance, and combined exercise training [ 15 , 16 ]. Moreover, metformin may have no association with osteoarthritis [ 26 ]. Our study found that diabetic patients treated with metformin had better physical functions, and a lower risk of osteoarthritis than non-metformin patients. However, subgroup analyses in different frailty groups failed to determine the effects of metformin on grip strength and walking speed. Metformin was associated with a reduced risk of osteoarthritis only in non-frail patients with diabetes. Different from the opinions of Laksmi et al., our study suggested that metformin treatment improved the quality of life (assessed by the SF-36) in both frail and non-frail patients with diabetes. As the relationship between metformin use and physical function is controversial, further investigations are required [ 14 ]. MET-PREVENT is an ongoing clinical double-blind trial aiming to explore the beneficial effects of metformin on physical performance in older people with sarcopenia and physical prefrailty/frailty, which uses the 4 m walk speed at the 4-month follow-up visit as the primary outcome [ 27 ].

Metformin is a strongly recommended first-line drug for patients with type 2 diabetes, especially for those with obesity. Our findings showed that high BMI and high levels of nutrient metabolism indicators, such as blood glucose, blood lipids, and proteins, and low levels of kidney function indicators, D-dimer, and NT-proBNP were associated with metformin in diabetic patients, especially in non-frail patients. Insulin resistance and inflammation, which are common pathological processes in frailty and diabetes, are also targets of metformin. Therefore, metformin may be a promising drug for the prevention of frailty. The potential mechanisms of metformin mainly include the upregulation of nutrient-sensing pathways by activating AMP-activated kinase (AMPK), inhibiting mTORC1 [ 28 ], as well as downregulation of insulin/insulin-like growth factor-1 (IGF-1) signaling. Moreover, metformin enhances autophagy and maintains mitochondrial function to reduce age-related inflammation and reactive oxygen species [ 29 ]. A recent study found that metformin prevented female mice from age-associated ovarian fibrosis by modulating specific immune cells and fibroblasts [ 30 ].

The Targeting Aging with Metformin (TAME) clinical trial reported eight serum biomarkers associated with geoscience, including interleukin-6, CRP, tumor necrosis factor receptor II, growth differentiation factor -15, insulin, IGF-1, cystatin C, NT-proBNP, and hemoglobin A1c [ 31 ]. In the present study, high fasting plasma glucose, hemoglobin A1c, triglyceride, prealbumin, albumin, hemoglobin, urine mALB/Cre levels and low creatine, blood urea nitrogen, D-dimer, and NT-proBNP levels were associated with metformin. Low levels of CRP were only observed in frail patients with diabetes. Further comparisons found differences between serum NT-proBNP and D-dimer levels and metformin treatment. D-dimer might play a particularly important role in frailty [ 32 , 33 ]. As the Cardiovascular Healthy Study mentioned earlier, high levels of D-dimer were observed in frailty older adults even after adjusting for cardiovascular disease and diabetes [ 33 ]. Those with high interleukin-6 and high D-dimer levels had the greatest functional declines [ 32 ]. Higher D-dimer and high tissue plasminogen activator was associated with an increased risk of frailty (OR = 2.20 95%CI 1.29–3.75) [ 34 ]. Cross-sectional analysis evidence indicated that elevated plasma BNP was associated with increased risks of frailty, prefrailty, and low levels of grip strength and walking speed in older adults [ 35 ]. Moreover, NT-proBNP has been identified as a simple and useful tool for frailty assessment in patients with newly diagnosed multiple myeloma [ 36 ]. High NT-proBNP levels are also strongly associated with incident frailty in the community-dwelling population without known cardiovascular disease [ 37 ]. Future studies are needed to verify the feasibility, effectiveness, predictability, and responsiveness of these serum biomarkers, as well as explore novel biomarkers.

Metformin was associated with a reduction of all-cause mortality and age-related diseases [ 10 , 38 ]. Clinical studies have shown that metformin has protective effects on coronary atherosclerotic heart disease in patients with diabetes [ 39 , 40 ] and coronary atherosclerosis in males with prediabetes and early diabetes [ 41 ]. A retrospective 5-year follow-up cohort study on Chinese adults with diabetes also showed that metformin treatment could reduce the risk of coronary atherosclerotic heart disease events and all-cause mortality [ 42 ]. The Glucose Lowering In Non-diabetic hyperglycemia Trial (GLINT) also found benefits of metformin for coronary atherosclerotic heart disease in non-diabetic hyperglycemia adults [ 43 ]. The follow-up results showed that metformin use in patients with diabetes was in accordance with a lower risk of adverse outcomes. Frail diabetic patients were related to high risk of adverse outcomes, including cardiovascular events, hospitalization, death, and combined primary outcomes. Consistent with the above studies, our study also found that metformin-treated patients had a significantly lower risk of cardiovascular events than non-treated patients, regardless of frailty. However, no effect of metformin on cerebrovascular disease, hospitalization, or death was observed in the subgroup analyses. Therefore, the protective effect of metformin on coronary atherosclerotic heart disease in patients with diabetes may be more sensitive than that of other adverse events. The Cox regression analysis indicated that metformin was an independent protective factor and frailty was an independent risk factor for adverse events. The protective effect of metformin on the combined primary outcomes was observed only in non-frail patients. Thus, the protective effect of metformin against adverse outcomes may be affected by the degree of frailty. One of the potential reasons is that frail diabetic patients were older and had a lower proportion of metformin treatment than that non-metformin (Table 1 ), which may be influenced by the disease state, the tolerance of metformin gastrointestinal adverse reactions, and the metformin contraindications. In addition, controversy exists regarding the beneficial effects of metformin on physical function. As multiple physiological changes from cellular/molecular level, system level, to organ level contribute to frailty [ 44 ]. Recently, an integrated care model is recommended for frailty management, which includes physical exercise, nutrition, comorbidity and polypharmacy management, social support, etc. [ 45 , 46 ]. While research on the pharmacological treatment of frailty is still in the exploratory stage. Metformin treatment has a weaker effect on frailty, and comprehensive assessment and management are necessary for frailty management. As frailty is a dynamic process, and early frailty is reversible. The protective effect of metformin in non-frail patients may help prevent frailty progression and adverse outcomes.

There may be some possible limitations in this study. First, the sample size was small, and all participants were from a single center. There may exhibit a certain degree of selection bias. For example, frail patients treated with metformin were 13.1%, lower than that not treated with metformin (28.2%). The lower frailty proportion in the metformin group may be also influenced by disease conditions and metformin contraindications. Frailty is generally comorbid with cardiac failure, chronic kidney disease, hepatic failure and cirrhosis, and respiratory insufficiency, such as chronic obstructive diseases, which are recognized as metformin contraindications [ 47 ]. This may decrease the application of metformin in frail population. Second, comprehensive assessment and laboratory tests were not performed during the follow-up. Although we found a negative association between metformin and frailty, the clinical value was inferior to the findings from randomized controlled trials. Future research requires prospective clinical studies and validation in larger populations.

Conclusion

In conclusion, metformin use is independently associated with a low risk of frailty in patients with diabetes. Metformin had an independent protective effect on adverse events, whereas frailty was an independent risk factor for worse outcomes. The benefits of metformin use were observed in non-frail patients with diabetes in this study, but not in those with frailty. Thus, frailty may weaken the long-term protective effects of metformin. Early identification and timely intervention of the frail state of patients with diabetes may help enhance the benefits of metformin. However, further prospective clinical studies are needed to verify these findings.

Supplementary Information

Additional file 1: Table S1. Univariate regression analysis of risk factors for frailty.

Abbreviations

Body mass index

C-reactive protein

N-terminal pro-B-type natriuretic peptide

Microalbuminuria/creatinine

AMP-activated kinase

Insulin/insulin-like growth factor-1

Targeting Aging with Metformin

Glucose lowering in non-diabetic hyperglycemia trial

Pan Liu and Yiming Pan as the co-first authors of this article.

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Acknowledgements

We acknowledge Drs. Bixi Li, Yaxin Zhang, Shijie Li, Tong Ji, Yumeng Chen, Ying Li, Ou Zhao, Shijie Li, Guanzhen Wang, and Li Zhang for their support in data collection for this study.

Author contributions

LM and YL designed the study. PL and YP were major contributors in writing the manuscript. YS, YZ, WZ, XL, JL collected, analyzed, and interpreted the data. All authors read and approved the final manuscript.

Funding

This work was supported by the National Key R&D Program of China (2020YFC2008604).

Availability of data and materials

All data published here are under the consent for publication.

Declarations

Ethics approval and consent to participate

This study was approved by the ethics committee of Xuanwu Hospital Capital Medical University. Information was only collected after written consent was obtained from all participants. This study was conducted in accordance with the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

All authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as potential competing interests.

References

1. International Diabetes Federation. IDF Diabetes Atlas, 10th edn. Brussels, Belgium; 2021, Nov, 02. https://www.idf.org/news/240:diabetes-now-affects-one-in-10-adults-worldwide.html .

2. Li Y Teng D Shi X Prevalence of diabetes recorded in mainland China using 2018 diagnostic criteria from the American diabetes association: national cross sectional study BMJ 2020 369 m997 32345662 10.1136/bmj.m997 PMC7186854

3. Fried LP Tangen CM Walston J Frailty in older adults: evidence for a phenotype J Gerontol A Biol Sci Med Sci 2001 56 3 M146 156 11253156 10.1093/gerona/56.3.m146

4. Clegg A Young J Iliffe S Rikkert MO Rockwood K Frailty in elderly people Lancet 2013 381 9868 752 762 23395245 10.1016/S0140-6736(12)62167-9 PMC4098658

5. Chhetri JK Zheng Z Xu X Ma C Chan P The prevalence and incidence of frailty in Pre-diabetic and diabetic community-dwelling older population: results from Beijing longitudinal study of aging II (BLSA-II) BMC Geriatr 2017 17 1 47 28178934 10.1186/s12877-017-0439-y PMC5299771

6. Liccini A Malmstrom TK Frailty and sarcopenia as predictors of adverse health outcomes in persons with diabetes mellitus J Am Med Dir Assoc 2016 17 9 846 851 27569712 10.1016/j.jamda.2016.07.007

7. Park SW Goodpaster BH Strotmeyer ES Accelerated loss of skeletal muscle strength in older adults with type 2 diabetes: the health, aging, and body composition study Diabetes Care 2007 30 6 1507 1512 17363749 10.2337/dc06-2537

8. Morley JE Malmstrom TK Rodriguez-Mañas L Sinclair AJ Frailty, sarcopenia and diabetes J Am Med Dir Assoc 2014 15 12 853 859 25455530 10.1016/j.jamda.2014.10.001

9. Mohammed I Hollenberg MD Ding H Triggle CR A critical review of the evidence that metformin is a putative anti-aging drug that enhances healthspan and extends lifespan Front Endocrinol 2021 12 718942 10.3389/fendo.2021.718942 PMC8374068 34421827

10. Campbell JM Bellman SM Stephenson MD Lisy K Metformin reduces all-cause mortality and diseases of ageing independent of its effect on diabetes control: a systematic review and meta-analysis Ageing Res Rev 2017 40 31 44 28802803 10.1016/j.arr.2017.08.003

11. Kulkarni AS Gubbi S Barzilai N Benefits of metformin in attenuating the hallmarks of aging Cell Metab 2020 32 1 15 30 32333835 10.1016/j.cmet.2020.04.001 PMC7347426

12. Sumantri S Setiati S Purnamasari D Dewiasty E Relationship between metformin and frailty syndrome in elderly people with type 2 diabetes Acta Med Indones 2014 46 3 183 188 25348180

13. Wang C-P Lorenzo C Habib SL Jo B Espinoza SE Differential effects of metformin on age related comorbidities in older men with type 2 diabetes J Diabetes Complicat 2017 31 4 679 686 10.1016/j.jdiacomp.2017.01.013 PMC5654524 28190681

14. Laksmi PW Setiati S Tamin TZ Effect of metformin on handgrip strength, gait speed, Myostatin serum level, and health-related quality of life: a double blind randomized controlled trial among non-diabetic pre-frail elderly patients Acta Med Indones 2017 49 2 118 127 28790226

15. Walton RG Dungan CM Long DE Metformin blunts muscle hypertrophy in response to progressive resistance exercise training in older adults: a randomized, double-blind, placebo-controlled, multicenter trial: The MASTERS trial Aging Cell 2019 18 6 e13039 31557380 10.1111/acel.13039 PMC6826125

16. Konopka AR Miller BF Taming expectations of metformin as a treatment to extend healthspan GeroScience 2019 41 2 101 108 30746605 10.1007/s11357-019-00057-3 PMC6544683

17. Ma L Zhang L Sun F Li Y Tang Z Cognitive function in Prefrail and frail community-dwelling older adults in China BMC Geriatr 2019 19 1 53 30813907 10.1186/s12877-019-1056-8 PMC6391822

18. Wu C Smit E Xue QL Odden MC Prevalence and correlates of frailty among community-dwelling Chinese older adults: the China health and retirement longitudinal study J Gerontol A Biol Sci Med Sci 2017 73 1 102 108 28525586 10.1093/gerona/glx098 PMC5861883

19. Radloff LS The CES-D scale: a self-report depression scale for research in the general population Appl Psychol Meas 1977 1 3 385 401

20. Li L Wang H Shen Y Development and psychometric tests of a Chinese version of the SF-36 Health Survey Scales Chinese J Prev Med 2002 36 2 109 113 12410965

21. Ware JE Jr Sherbourne CD The MOS 36-ltem short-form health survey (SF-36): I. conceptual framework and item selection Med Care 1992 30 6 473 483 1593914

22. Brazier J Roberts J Deverill M The estimation of a preference-based measure of health from the SF-36 J Health Econ 2002 21 2 271 292 11939242 10.1016/s0167-6296(01)00130-8

23. Wang C-P Lorenzo C Espinoza SE Frailty attenuates the impact of metformin on reducing mortality in older adults with type 2 diabetes J Endocrinol Diabetes Obes 2014 2 2 1031 25506599 PMC4264048

24. Dujic T Zhou K Donnelly LA Tavendale R Palmer CN Pearson ER Association of organic cation transporter 1 with intolerance to metformin in type 2 diabetes: a GoDARTS study Diabetes 2015 64 5 1786 1793 25510240 10.2337/db14-1388 PMC4452716

25. Baskaran D Aparicio-Ugarriza R Ferri-Guerra J Milyani R Florez H Ruiz JG Is there an association between metformin exposure and frailty? Gerontol Geriatr Med 2020 6 2333721420924956 32596419 10.1177/2333721420924956 PMC7297486

26. Barnett LA Jordan KP Edwards JJ Windt DA Does metformin protect against osteoarthritis? An electronic health record cohort study Prim Health Care Res Dev 2017 18 6 623 628 28539134 10.1017/S1463423617000287

27. Rennie KJ Witham M Bradley P MET-PREVENT: metformin to improve physical performance in older people with sarcopenia and physical prefrailty/frailty - protocol for a double-blind, randomised controlled proof-of-concept trial BMJ Open 2022 12 7 e061823 10.1136/bmjopen-2022-061823 PMC9297211 35851031

28. Howell JJ Hellberg K Turner M Metformin inhibits hepatic mTORC1 signaling via dose-dependent mechanisms involving AMPK and the TSC complex Cell Metab 2017 25 2 463 471 28089566 10.1016/j.cmet.2016.12.009 PMC5299044

29. Bharath LP Agrawal M McCambridge G Metformin enhances autophagy and normalizes mitochondrial function to alleviate aging-associated inflammation Cell Metab 2020 32 1 44 55.e46 32402267 10.1016/j.cmet.2020.04.015 PMC7217133

30. Landry DA Yakubovich E Cook DP Fasih S Upham J Vanderhyden BC Metformin prevents age-associated ovarian fibrosis by modulating the immune landscape in female mice Sci Adv 2022 8 35 eabq1475 36054356 10.1126/sciadv.abq1475 PMC10848964

31. Justice JN Ferrucci L Newman AB A framework for selection of blood-based biomarkers for geroscience-guided clinical trials: report from the tame biomarkers workgroup GeroScience 2018 40 5–6 419 436 30151729 10.1007/s11357-018-0042-y PMC6294728

32. Cohen HJ Harris T Pieper CF Coagulation and activation of inflammatory pathways in the development of functional decline and mortality in the elderly Am J Med 2003 114 3 180 187 12637131 10.1016/s0002-9343(02)01484-5

33. Walston J McBurnie MA Newman A Frailty and activation of the inflammation and coagulation systems with and without clinical comorbidities: results from the cardiovascular health study Arch Intern Med 2002 162 20 2333 2341 12418947 10.1001/archinte.162.20.2333

34. Reiner AP Aragaki AK Gray SL Inflammation and thrombosis biomarkers and incident frailty in postmenopausal women Am J Med 2009 122 10 947 954 19682668 10.1016/j.amjmed.2009.04.016 PMC2754604

35. Yao S Guo J Shi G Association of BNP with frailty in elderly population: rugao longevity and ageing study J Nutr Health Aging 2019 23 1 73 78 30569072 10.1007/s12603-018-1112-7

36. Milani P Vincent Rajkumar S Merlini G N-terminal fragment of the type-B natriuretic peptide (NT-proBNP) contributes to a simple new frailty score in patients with newly diagnosed multiple myeloma Am J Hematol 2016 91 11 1129 1134 27508522 10.1002/ajh.24532 PMC5129512

37. Jia Y Li D Yu J Subclinical cardiovascular disease and frailty risk: the atherosclerosis risk in communities study BMC Geriatr 2022 22 1 321 35413794 10.1186/s12877-022-02974-z PMC9006603

38. Ekström N Schiöler L Svensson A-M Effectiveness and safety of metformin in 51 675 patients with type 2 diabetes and different levels of renal function: a cohort study from the swedish national diabetes register BMJ Open 2012 10.1136/bmjopen-2012-001076 PMC3400073 22798258

39. Müller-Werdan U Cardiovascular prevention in old age-cardiovascular prevention of ageing? Z Gerontol Geriatr 2022 55 6 461 464 36044072 10.1007/s00391-022-02111-w

40. Paul SK Klein K Majeed A Khunti K Association of smoking and concomitant metformin use with cardiovascular events and mortality in people newly diagnosed with type 2 diabetes J Diabetes 2016 8 3 354 362 25929583 10.1111/1753-0407.12302

41. Goldberg RB Aroda VR Bluemke DA Effect of long-term metformin and lifestyle in the diabetes prevention program and its outcome study on coronary artery calcium Circulation 2017 136 1 52 64 28476766 10.1161/CIRCULATIONAHA.116.025483 PMC5526695

42. Fung CSC Wan EYF Wong CKH Jiao F Chan AKC Effect of metformin monotherapy on cardiovascular diseases and mortality: a retrospective cohort study on Chinese type 2 diabetes mellitus patients Cardiovasc Diabetol 2015 14 137 26453464 10.1186/s12933-015-0304-2 PMC4600251

43. Griffin SJ Bethel MA Holman RR Metformin in non-diabetic hyperglycaemia: the GLINT feasibility RCT Health Technol Assess 2018 22 18 1 64 10.3310/hta22180 PMC5925436 29652246

44. Fried LP Cohen AA Xue Q-L Walston J Bandeen-Roche K Varadhan R The physical frailty syndrome as a transition from homeostatic symphony to cacophony Nat Aging 2021 1 1 36 46 34476409 10.1038/s43587-020-00017-z PMC8409463

45. Ruiz JG Dent E Morley JE Screening for and managing the person with frailty in primary care: ICFSR consensus guidelines J Nutr Health Aging 2020 24 9 920 927 33155616 10.1007/s12603-020-1498-x PMC7597431

46. Dent E Morley JE Cruz-Jentoft AJ Physical frailty: ICFSR international clinical practice guidelines for identification and management J Nutr Health Aging 2019 23 9 771 787 31641726 10.1007/s12603-019-1273-z PMC6800406

47. Roxana Adriana S Diana Simona Ș Manfredi R Manfredi R Anca Mihaela Pantea S Metformin indications, dosage, adverse reactions, and contraindications Metformin 2020 Rijeka IntechOpen