ABSTRACT
Aims/Introduction
We aimed to clarify the real‐world risk of lower‐limb amputation and identify factors related to increased risk in Japanese patients with type 2 diabetes using sodium–glucose cotransporter 2 inhibitors (SGLT2is).
Materials and Methods
We carried out a retrospective observational cohort study utilizing the Japanese Medical Data Vision, a diagnosis procedure combination database. We identified 107,296 patients with type 2 diabetes who were initiated on SGLT2is or metformin (control; n = 53,648 per group) using 1:1 propensity score matching from April 2014 to October 2019. The hazard ratio (HR) for the risk of lower‐limb amputation was analyzed using a Cox proportional hazards model adjusted for patients' baseline characteristics and use of concomitant medical agents.
Results
Of the 107,296 patients, 66 (0.06%); that is, 41 (0.08%) in the SGLT2is group and 25 (0.05%) in the metformin group, underwent amputation, with no significant difference in the proportions between the groups. There was no significant difference in the risk of amputation between the SGLT2is and metformin groups (HR 1.34, 95% confidence interval [CI] 0.80–2.24). However, female sex (HR 2.78, 95% CI 1.12–6.94) and use of strong statins (HR 2.68; 95% CI 1.18–8.20) were significantly associated with a higher risk of amputation in the SGLT2is group than in the metformin group.
Conclusions
SGLT2is might not be related to an increased risk of lower‐limb amputation in patients with type 2 diabetes in real‐world clinical practice. The possible increased risk of SGLT2is‐associated amputation in female patients with type 2 diabetes and patients with type 2 diabetes requiring strong statins is notable.
Keywords: Lower‐limb amputation, Sodium–glucose cotransporter 2 inhibitor, Type 2 diabetes
We aimed to clarify the real‐world risk of lower‐limb amputation and identify factors associated with increased risk in Japanese patients with type 2 diabetes using sodium–glucose cotransporter 2 inhibitors (SGLT2is). Of the 107,296 patients, 66 (0.06%); that is, 41 (0.08%) in the SGLT2is group and 25 (0.05%) in the metformin group, underwent amputation, with no significant difference in the proportions between the groups. Although there was no significant difference in the risk of amputation between the SGLT2is and metformin groups (hazard ratio 1.34, 95% confidence interval 0.80–2.24), female sex (hazard ratio: 2.78; 95% confidence interval 1.12–6.94) and use of strong statins (hazard ratio 2.68; 95% confidence interval 1.18–8.20) were significantly associated with a higher risk of amputation in the SGLT2is group than in the metformin group.
INTRODUCTION
Type 2 diabetes is a metabolic disorder characterized by elevated levels of plasma glucose. Hyperglycemia is related to an increase in insulin resistance and a decrease in β‐cell function, and long‐standing hyperglycemia can cause microvascular and macrovascular complications. Lower‐limb amputation is a serious outcome in patients with type 2 diabetes 1 . In particular, amputation risk is reportedly increased in patients with poorly controlled and long‐standing type 2 diabetes, especially in those requiring insulin therapy 2 . In Japan, the amputation rate in patients with diabetes and peripheral artery disease (PAD) was reportedly 4.43‐fold higher than that in patients without diabetes having PAD alone 3 .
Currently, sodium–glucose cotransporter 2 inhibitors (SGLT2is) are popularly prescribed to patients with type 2 diabetes. SGLT2is ameliorate hyperglycemia through inhibiting the renal proximal tubular reabsorption of glucose, independently of insulin secretion 4 . This distinct class of oral hypoglycemic agents can reduce glycated hemoglobin levels by 0.6–0.8% without increasing the risk of hypoglycemia, thereby increasing time in range, reducing bodyweight and improving metabolic factors, including blood pressure and lipid profile 5 . In contrast, the CANagliflozin cardioVascular Assessment Study (CANVAS) Program showed an approximately twofold incidence of lower‐limb amputation in patients with type 2 diabetes treated with canagliflozin than that in those treated with placebo (6.3 vs 3.4 participants per 1,000 patient‐years) 6 .
The mechanism for the risk of SGLT2is‐associated amputation is unclear 7 ; this adverse event might be associated with volume depletion, leading to circulation insufficiency in the peripheral vasculature 6 , 8 , raising concerns about the risk of amputations associated with canagliflozin and other SGLT2is. Systematic reviews and meta‐analyses using randomized controlled trials or observational studies have shown that SGLT2is are not related to the increased risk of amputation 9 , 10 , 11 . However, whether the findings are true in real‐world settings in Japan or not remains unclear.
Thus, the present study aimed to show the actual risk of lower‐limb amputation in patients with type 2 diabetes using SGLT2is compared with that in those using metformin as a control in the real‐world use of SGLT2is in Japan, and to define the association between amputation risk and patients' clinical characteristics.
MATERIALS AND METHODS
Study design
This was a retrospective cohort study carried out utilizing the Medical Data Vision (MDV) administrative claims database, a nationwide database using data from Japanese acute care hospitals utilizing the Diagnosis Procedure Combination (DPC) system. The date when SGLT2is or metformin was first prescribed to patients in the SGLT2is and metformin groups, respectively, was defined as the observation start date. Patients were censored at the timing of undergoing lower‐limb amputation; discontinuation of target drugs (SGLT2is or metformin); addition of metformin or SGLT2is administration in patients in the SGLT2is or metformin groups, respectively; loss of insurance coverage; or death. The minimum observation period was 1 day after the initiation of SGLT2is or metformin.
Study population
First, we extracted data of 3,129,105 patients with diabetes mellitus (International Classification of Diseases, 10th revision [ICD‐10] code: E11–E14) registered between 1 April 2008 and 31 October 2019 in the MDV database (Figure 1). Next, we identified patients with type 2 diabetes aged ≥18 years who were newly prescribed SGLT2is and/or metformin after April 2014, the month in which SGLT2i (ipragliflozin) was available for the first time in Japan. The exclusion criteria were as follows: (i) patients diagnosed with type 1 diabetes (ICD‐10 code: E10; n = 66,760), those with both type 1 diabetes and type 2 diabetes (ICD‐10 code: E11 or E14; n = 179,550), and those with any diabetes mellitus, except type 1 diabetes and type 2 diabetes (n = 860,024); (ii) patients aged <18 years and/or patients with no registration history of SGLT2is and metformin use (n = 1,525,828); and, (iii) patients prescribed metformin before April 2014 (n = 170,778). As a result, 326,165 patients with type 2 diabetes were eligible for inclusion and were subsequently divided into two groups: (i) new users of SGLT2is without preceding metformin use (n = 109,660); and (ii) new users of metformin without preceding SGLT2is use (n = 216,505). Next, we carried out propensity score matching to balance the baseline characteristics (age, sex, body mass index [BMI], and use of concomitant drugs including antidiabetics, antiplatelets, anticoagulants and statins) between the two groups in a 1:1 ratio for SGLT2is to metformin. Finally, a total of 107,296 patients, comprising new users of SGLT2is (n = 53,648) and metformin (n = 53,648), were eligible for inclusion.
Figure 1.
Patient disposition. SGLT2is, sodium–glucose cotransporter 2 inhibitors; T1D, type 1 diabetes; T2D, type 2 diabetes.
Identification of lower‐limb amputation events and use of concomitant drugs
Lower‐limb amputation events were identified using the relevant ICD‐10 codes; that is, Z89.4, Z89.5 and Z89.6. Concomitant drugs were those confirmed to have been prescribed for at least 30 days before SGLT2is or metformin initiation. Atorvastatin, pitavastatin and rosuvastatin were classified as strong statins, whereas other statins were classified as standard statins in the present study.
Outcome
The major outcome was to compare the risks of lower‐limb amputation between patients with type 2 diabetes using SGLT2is and metformin. Furthermore, the clinical parameters associated with the increased risk of amputation in patients with type 2 diabetes using SGLT2is and metformin were determined.
Patient characteristics
Age, sex and BMI were identified using data from the claim records within 30 days of the amputation date. Obesity was defined as a BMI of ≥25 kg/m2 12 .
Statistical analysis
Normally distributed data (age and BMI) are presented as mean ± standard deviation. Continuous variables were analyzed using the unpaired t‐test. Categorical variables were analyzed using the χ2‐test or Fisher's exact test, and are expressed as absolute numbers or percentages. As aforementioned, to minimize or eliminate potential confounding effects, we adjusted for differences in baseline characteristics by 1:1 propensity score matching using the nearest neighbor matching method 13 . After matching, the hazard ratio (HR) for amputation was analyzed using a Cox proportional hazards model. All statistical analyses were carried out using IBM SPSS Statistics for Windows, version 23.0 (IBM Corp., Armonk, NY, USA). Differences were considered statistically significant at a P‐value of <0.05.
RESULTS
Table S1 and Table 1 present the baseline clinical characteristics before and after 1:1 propensity score matching, respectively. Before propensity score matching, there were significant differences in many clinical parameters between the SGLT2is and metformin groups, making it difficult to statistically investigate the effect of SGLT2is on amputation (Table S1). After propensity score matching, a total of 107,296 patients (n = 53,648 per group) were selected for the analysis (Table 1).
Table 1.
Baseline clinical characteristics of patients with type 2 diabetes using metformin or sodium–glucose cotransporter 2 inhibitors after 1:1 propensity score matching and the number of those who underwent lower‐limb amputation during the observation period
| Characteristics | Overall (n = 107,296) | Metformin (n = 53,648) | SGLT2is (n = 53,648) | P‐value (SGLT2is vs metformin) | |
|---|---|---|---|---|---|
| Sex, n (%) | Male | 70,013 (65.3) | 35,057 (65.3) | 34,956 (65.2) | 0.521† |
| Female | 37,283 (34.7) | 18,591 (34.7) | 18,692 (34.8) | ||
| Age | Mean (years) | 64.4 ± 12.8 | 64.4 ± 12.2 | 64.5 ± 13.4 | 0.227‡ |
| Distribution, n (%) | |||||
| <25 | 448 (0.4) | 209 (0.4) | 239 (0.4) | <0.001† | |
| 25–44 | 8,286 (7.7) | 3,986 (7.4) | 4,300 (8.0) | ||
| 45–64 | 39,834 (37.1) | 20,733 (38.6) | 19,101 (35.6) | ||
| ≥65 | 58,724 (54.7) | 28,720 (53.5) | 30,004 (55.9) | ||
| Body mass index | Mean (kg/m2) | 26.7 ± 5.1 | 26.7 ± 5.1 | 26.8 ± 5.2 | 0.134‡ |
| Distribution, n (%) | |||||
| <25 | 43,600 (40.6) | 21,762 (40.6) | 21,838 (40.7) | 0.639† | |
| ≥25 | 63,695 (59.4) | 31,886 (59.4) | 31,809 (59.3) | ||
| Amputation during the observation period | Distribution, n (%) | ||||
| No | 107,230 (99.94) | 53,623 (99.95) | 53,607 (99.92) | 0.065† | |
| Yes | 66 (0.06) | 25 (0.05) | 41 (0.08) | ||
| Breakdown of SGLT2is | |||||
| Canagliflozin | – | – | 8,402 (15.7) | – | |
| Dapagliflozin | – | – | 10,390 (19.4) | – | |
| Empagliflozin | – | – | 16,231 (30.3) | – | |
| Ipragliflozin | – | – | 9,900 (18.5) | – | |
| Luseogliflozin | – | – | 3,600 (6.7) | – | |
| Tofogliflozin | – | – | 5,125 (9.6) | – | |
| Other anti‐diabetic agents | |||||
| Insulin | Distribution, n (%) | 34,603 (32.3) | 17,277 (32.2) | 17,326 (32.3) | 0.754† |
| Thiazolidinediones | 7,391 (6.9) | 3,798 (7.1) | 3,593 (6.7) | 0.014† | |
| Sulfonylureas | 23,563 (22.0) | 11,731 (21.9) | 11,832 (22.1) | 0.461† | |
| Glinides | 9,758 (9.1) | 4,811 (9.0) | 4,947 (9.2) | 0.152† | |
| α‐GIs | 11,207 (10.4) | 5,574 (10.4) | 5,633 (10.5) | 0.563† | |
| DPP‐4 inhibitors | 63,779 (59.4) | 32,027 (59.7) | 31,752 (59.2) | 0.089† | |
| GLP‐1 receptor agonists | 4,523 (4.2) | 2,198 (4.1) | 2,325 (4.3) | 0.056† | |
| Antiplatelet/anticoagulant drugs | |||||
| Aspirin | Distribution, n (%) | 22,942 (21.4) | 11,521 (21.5) | 11,421 (21.3) | 0.461† |
| P2Y12 inhibitors | 16,067 (15.0) | 8,011 (14.9) | 8,056 (15.0) | 0.707† | |
| PDE inhibitors | 3,945 (3.7) | 1,991 (3.7) | 1,954 (3.6) | 0.559† | |
| Prostaglandins | 2,242 (2.1) | 1,153 (2.1) | 1,089 (2.0) | 0.179† | |
| 5HT2 inhibitors | 492 (0.5) | 265 (0.5) | 227 (0.4) | 0.095† | |
| Warfarin | 3,420 (3.2) | 1,715 (3.2) | 1,705 (3.2) | 0.876† | |
| DOACs | 6,331 (5.9) | 3,181 (5.9) | 3,150 (5.9) | 0.698† | |
| Statins | |||||
| Standard statins | Distribution, n (%) | 5,138 (4.8) | 2,635 (4.9) | 2,503 (4.7) | 0.061† |
| Strong statins | 38,256 (35.7) | 19,099 (35.6) | 19,157 (35.7) | 0.716† | |
Data are shown as the mean ± standard deviation. Body mass index is weight in kilograms divided by the square of height in meters. †χ2‐test; ‡unpaired t‐test. α‐GI, α‐glucosidase inhibitor; DOAC, direct oral anticoagulant; DPP4, dipeptidyl peptidase‐4; 5HT2, 5‐hydroxytryptamine 2; GLP‐1, glucagon‐like peptide‐1; PDE, phosphodiesterase; SGLT2i, sodium–glucose cotransporter 2 inhibitor.
Lower‐limb amputation was carried out in 66 (0.06%) of the 107,296 patients with type 2 diabetes; that is, 25 (0.05%) of 53,648 in the metformin group and 41 (0.08%) of 53,648 in the SGLT2is group (Table 1), over a mean observation period of 142.8 ± 145.5 days. There was no significant difference in the incidence of amputation between the groups, although the incidence tended to be higher in the SGLT2is group than in the metformin group (P = 0.065, χ2‐test; Table 1). Furthermore, there were no significant between‐group differences in sex, mean age and BMI. In addition, there were no significant differences in the usage rates of all anti‐diabetic drugs, excluding thiazolidinediones, antiplatelet and anticoagulant drugs, and statins. These findings suggest that most confounding effects could be eliminated, leading to a reliable analysis of the association between SGLT2is use and lower‐limb amputation. In the present study, the most frequently used SGLT2i was empagliflozin (30.3%), and the frequencies of the use of the other SGLT2is decreased in the following order: dapagliflozin (19.4%), ipragliflozin (18.5%), canagliflozin (15.7%), tofogliflozin (9.6%) and luseogliflozin (6.7%; Table 1).
The clinical characteristics of patients with type 2 diabetes undergoing lower‐limb amputation, divided into the metformin and SGLT2is groups, after 1:1 propensity score matching, are presented in Table 2. The mean age and BMI were significantly higher in the SGLT2is group than in the metformin group. In addition, the proportion of obese patients was significantly higher in the SGLT2is group than in the metformin group.
Table 2.
Clinical characteristics of patients with type 2 diabetes undergoing lower‐limb amputation divided into metformin and sodium–glucose cotransporter 2 inhibitors groups after propensity score matching
| Characteristics | Overall (n = 66) | Metformin (n = 25) | SGLT2is (n = 41) | P‐value (SGLT2is vs metformin) | |
|---|---|---|---|---|---|
| Sex, n (%) | Male | 39 (59.1) | 19 (76.0) | 20 (48.8) | 0.054† |
| Female | 27 (40.9) | 6 (24.0) | 21 (51.2) | ||
| Age | Mean (years) | 64.5 ± 13.7 | 60.1 ± 13.0 | 67.1 ± 13.5 | 0.021§ |
| Distribution, n (%) | |||||
| <25 | 0 (0) | 0 (0) | 0 (0) | 0.019‡ | |
| 25–44 | 8 (12.1) | 5 (20.0) | 3 (7.3) | ||
| 45–64 | 20 (30.3) | 11 (44.0) | 9 (22.0) | ||
| ≥65 | 38 (57.6) | 9 (36.0) | 29 (70.7) | ||
| Body mass index | Mean (kg/m2) | 24.0 ± 6.0 | 22.1 ± 2.2 | 25.2 ± 7.1 | 0.023§ |
| Distribution, n (%) | |||||
| <25 | 47 (71.2) | 25 (100) | 22 (53.7) | <0.001† | |
| ≥25 | 19 (28.8) | 0 (0) | 19 (46.3) | ||
| Breakdown of SGLT2is | |||||
| Canagliflozin | – | – | 6 | – | |
| Dapagliflozin | – | – | 10 | – | |
| Empagliflozin | – | – | 18 | – | |
| Ipragliflozin | – | – | 4 | – | |
| Luseogliflozin | – | – | 1 | – | |
| Tofogliflozin | – | – | 2 | – | |
| Other anti‐diabetic agents | |||||
| Insulin | Distribution, n (%) | 30 (45.5) | 15 (60.0) | 15 (36.6) | 0.110† |
| Thiazolidinediones | 5 (7.6) | 1 (4.0) | 4 (9.8) | 0.642‡ | |
| Sulfonylureas | 13 (19.7) | 6 (24.0) | 7 (17.1) | 0.535‡ | |
| Glinides | 4 (6.1) | 3 (12.0) | 1 (2.4) | 0.148‡ | |
| α‐GIs | 6 (9.1) | 1 (4.0) | 5 (12.2) | 0.396‡ | |
| DPP‐4 inhibitors | 39 (59.1) | 19 (76.0) | 20 (48.8) | 0.054† | |
| GLP‐1 receptor agonists | 4 (6.1) | 1 (4.0) | 3 (7.3) | 1.000‡ | |
| Antiplatelet/anticoagulant drugs | |||||
| Aspirin | Distribution, n (%) | 18 (27.3) | 8 (32.0) | 10 (24.4) | 0.698† |
| P2Y12 inhibitors | 6 (9.1) | 3 (12.0) | 3 (7.3) | 0.666‡ | |
| PDE inhibitors | 5 (7.6) | 1 (4.0) | 4 (9.8) | 0.642‡ | |
| Prostaglandins | 3 (4.5) | 2 (8.0) | 1 (2.4) | 0.552‡ | |
| 5HT2 inhibitors | 1 (1.5) | 0 (0) | 1 (2.4) | 1.000‡ | |
| Warfarin | 2 (3.0) | 1 (4.0) | 1 (2.4) | 1.000‡ | |
| DOACs | 8 (12.1) | 2 (8.0) | 6 (14.6) | 0.700‡ | |
| Statins | |||||
| Standard statins | Distribution, n (%) | 3 (4.5) | 2 (8.0) | 1 (2.4) | 0.552‡ |
| Strong statins | 18 (27.3) | 4 (16.0) | 14 (34.1) | 0.187† | |
Data are shown as mean ± standard deviation. Body mass index is the weight in kilograms divided by the square of height in meters. † χ2‐test, ‡Fisher's exact test, §unpaired t‐test. α‐GI, α‐glucosidase inhibitor; DOAC, direct oral anticoagulant; DPP4, dipeptidyl peptidase‐4; 5HT2, 5‐hydroxytryptamine 2; GLP‐1, glucagon‐like peptide‐1; PDE, phosphodiesterase; SGLT2i, sodium–glucose cotransporter 2 inhibitor.
Next, we compared the amputation risk associated with SGLT2is and metformin administration using the Cox proportional hazards model. There was no significant difference in the risk of lower‐limb amputation between the SGLT2is and metformin groups, although the risk in the former was seemingly higher than that in the latter (HR 1.34, 95% confidence interval [95% CI] 0.80–2.24; Figure 2). When analyzed according to each SGLT2i drug, there were no significant differences in the risks associated with each SGLT2i drug and metformin.
Figure 2.
Risk of lower‐limb amputation in patients with type 2 diabetes using sodium–glucose cotransporter 2 inhibitors (SGLT2is) compared with that in those using metformin. Forest plots show the hazard ratios (HRs) for lower‐limb amputation in patients with type 2 diabetes. Circles represent the HR, and horizontal bars extend from the lower limit to the upper limit of the 95% confidence interval (CI) of the estimated HR.
Subsequently, we compared the risks of lower‐limb amputation between the SGLT2is and metformin groups according to the baseline clinical characteristics. The risk was significantly higher in female patients using SGLT2is than in those using metformin (HR 2.78, 95% CI 1.12–6.94, P = 0.028; Figure 3); however, the same finding was not observed in male patients. Regarding age, the risk tended to be higher in patients aged ≥75 years using SGLT2is than in those using metformin (HR 2.32, 95% CI 0.88–6.13), although the difference was not significant (P = 0.090). Regarding BMI, although the HR for amputation could not be calculated in type 2 diabetes patients with a BMI of ≥25 kg/m2 due to the null event of amputation in those using metformin, the risk of lower‐limb amputation was considered higher in patients with BMI of ≥25 kg/m2 using SGLT2is than that in those using metformin. Meanwhile, the amputation risk was comparable between type 2 diabetes patients with a BMI of <25 kg/m2 using SGLT2is and those using metformin. Concomitant anti‐diabetic and antiplatelet/anticoagulant drugs did not influence amputation risk in the SGLT2is and metformin groups. Regarding statins, amputation risk was significantly higher in strong statin users using SGLT2is than in those using metformin (HR 2.68, 95% CI 1.18–8.20, P = 0.046).
Figure 3.
Comparison of the risk of lower‐limb amputation between patients with type 2 diabetes using sodium–glucose cotransporter 2 inhibitors (SGLT2is) and those using metformin, according to the baseline clinical characteristics. Forest plots show the hazard ratios (HRs) for lower‐limb amputation in patients with type 2 diabetes based on baseline clinical characteristics. Circles represent the HR, and horizontal bars extend from the lower limit to the upper limit of the 95% confidence interval (CI) of the estimated HR. α‐GI, α‐glucosidase inhibitor; BMI, body mass index; DOAC, direct oral anticoagulant; DPP4, dipeptidyl peptidase‐4; 5HT2, 5‐hydroxytryptamine 2; GLP‐1RA, glucagon‐like peptide‐1 receptor agonist; NA, not applicable; PDE, phosphodiesterase.
DISCUSSION
The present study showed no significant difference in the risk of lower‐limb amputation between patients with type 2 diabetes using SGLT2is and those using metformin. Considering that metformin was not reportedly related to an increased risk of amputation 2 , 14 , we concluded that SGLT2is were not related to an increased risk of lower‐limb amputation in patients with type 2 diabetes. However, a subanalysis according to sex using SGLT2is showed a 2.78‐fold increased risk of amputation compared with that using metformin in female patients with type 2 diabetes, but not in male patients. A similar finding was shown in a pooled analysis of the Empagliflozin Cardiovascular Outcome Event Trial in Type 2 Diabetes Mellitus Patients (EMPA‐REG OUTCOME), CANVAS Program, Dapagliflozin Effect on Cardiovascular Events–Thrombolysis in Myocardial Infarction 58 (DECLARE‐TIMI 58) and Canagliflozin and Renal Events in Diabetes with Established Nephropathy Clinical Evaluation (CREDENCE) trials; that is, the risk ratios for amputation in participants with SGLT2is treatment versus placebo were 1.79 (95% CI 1.08–2.97) and 1.23 (95% CI 0.99–1.52) in women and men, respectively, although there was no statistically significant difference between the two sexes 15 . However, the reason for the increased risk of amputation in women was not discussed and remains unclear. Generally, the hemodynamic effects of diuretic action and blood pressure lowering using SGLT2is are related to an increased risk of amputation 11 . Similar effects were found with diuretic agents, which might decrease peripheral perfusion and subsequently increase the risk of lower‐limb amputation in patients with type 2 diabetes 16 . Considering that, in general, the daily intake of water was reportedly lower in women than in men in Japan 17 and foreign countries 18 , female patients with type 2 diabetes might be more influenced by the hemodynamic effect of SGLT2is through hypovolemia due to insufficient water intake than male patients; however, this warrants further investigation. Furthermore, the mean age at which amputation was carried out was higher in female patients than in male patients; that is, 66.8 ± 14.5 years and 62.9 ± 12.9 years, respectively, although there was no significant difference between the two; this might be associated with the increased risk of amputation in female patients using SGLT2is presumably through a failure of local circulation due to advanced atherosclerosis in the foot. Nevertheless, the very small number of patients undergoing amputation and the lack of clinical information on possible risk factors for amputation, such as smoking status 19 and presence/absence of atherosclerotic vascular diseases 20 , might confound the reason for the sex difference in the amputation risk, which is a limitation of the present study.
A review of the published literature showed that risk factors for lower‐limb amputation associated with SGLT2is include concomitant cardiovascular diseases 20 . Furthermore, a recent meta‐analysis of randomized controlled trials in patients with type 2 diabetes showed that the risk of amputation increased in patients using SGLT2is, which was mainly driven by the results of cardiovascular outcome trials 11 . These findings suggest that atherosclerotic cardiovascular diseases might be closely related to SGLT2i‐associated amputation. In the present study, we could not collect patient information on the history of atherosclerotic cardiovascular disease, including PAD. However, considering that the risk of amputation was significantly higher in strong statin users using SGLT2is than in those using metformin in the present study, using SGLT2is in patients with type 2 diabetes at high risk of or affected by severe atherosclerosis requiring strong statins might be a risk factor for lower‐limb amputation.
Regarding obese patients with type 2 diabetes, amputation events were observed only in patients using SGLT2is, but not in those using metformin, although the reason for the null event in the metformin group remains unclear. To the best of our knowledge, few studies have shown the association between higher BMI and amputation risk in patients with type 2 diabetes using SGLT2is. Meanwhile, a recent meta‐analysis showed that diabetic foot ulcer‐related amputations are more likely to occur in patients with lower BMIs 21 . Diabetic foot ulcer, especially when concomitant with local infections, is a consumptive disease. A higher BMI might reflect better nutritional status and, hence, might work in favor of coping with a severe illness 22 . The mitigating effect of higher BMI on the risk of diabetic foot ulcer‐related amputations might be what is called “the obesity paradox,” a phenomenon involving better health outcomes associated with elevated bodyweight 23 , which could potentially explain the null event of amputation in obese patients with type 2 diabetes using metformin in the present study. However, using SGLT2is might counteract this phenomenon through the hemodynamic effect of diuresis. Alternatively, a recent meta‐analysis using data from randomized controlled trials showed that greater bodyweight reduction in SGLT2i users was associated with an increased risk of amputation 11 . The decrease in bodyweight associated with SGLT2is is reported to be greater in patients with type 2 diabetes having higher baseline BMIs 24 ; therefore, greater bodyweight reduction associated with SGLT2is might partly contribute to the increased number of amputation cases in obese patients with type 2 diabetes using SGLT2is in the present study.
The present study had several limitations that must be considered while interpreting the findings. First, limitations possibly stem from the use of diagnostic codes and prescription claims to identify patients with type 2 diabetes and ascertain amputation events. It is important to recognize that the MDV database does not represent the real conditions of all patients with diabetes, because it consists of patients treated mainly in acute care or emergency hospitals utilizing a DPC system, which mainly treat advanced‐stage patients. However, given that the proportions of patients with type 2 diabetes reported in the MDV database and the Japanese Health and Nutrition Examination Survey 25 were similar, the results of the present study might be applicable to most patients with type 2 diabetes. Second, the incidence of amputation was considerably low, possibly preventing an accurate analysis. In addition, considering that this was a retrospective cohort study, further prospective studies are necessary to evaluate the actual effect of SGLT2is on lower‐limb outcomes in the future. Third, we lacked important clinical information (glycated hemoglobin level 19 , diabetes duration 26 , smoking status 19 , diabetic microvascular complications, including diabetic retinopathy 19 , 27 , and macrovascular complications, including atherosclerotic diseases 20 ), all of which could be associated with amputation risk. In particular, it was difficult to determine whether patients developed complications or not, because the disease names registered in the DPC system usually included tentative names for the purpose of the diagnostic tests, which are hardly distinguishable from the actual disease names. Thus, we could not use information on complications in the present study. Furthermore, most of the lower‐limb amputations in patients with diabetes are often attributable to uncontrolled local infections and after necrosis of the extremities, which might be related to comorbidities, including diabetic foot and PAD 27 , 28 . Such conditions might precede surgical managements and ultimately lead to amputations. However, we could not obtain information regarding the presence/absence of infection, PAD or diabetic foot for the same reason stated above, possibly preventing an accurate analysis of the association between SGLT2is and lower‐limb amputation.
We attempted to carry out a re‐analysis by adding the information regarding hypertension and diabetic neuropathy, which are related to amputation 29 , 30 . Patients with hypertension were defined as those prescribed with any antihypertensive drugs, including renin–angiotensin system inhibitors. Patients with diabetic neuropathy were defined as those prescribed with tricyclic antidepressants, pregabalin and/or duloxetine. As a result, even after propensity score matching, statistically significant differences were noted in most baseline clinical characteristics between the two groups, distorting an accurate analysis on the effect of hypertension or diabetic neuropathy on the risk for lower‐limb amputation associated with SGLT2is, which is also a limitation of the present study.
In conclusion, SGLT2is might not be associated with an increased risk of lower‐limb amputation in patients with type 2 diabetes. However, female patients with type 2 diabetes using SGLT2is might have a higher risk of amputation than those using metformin. Furthermore, the amputation risk might be higher in patients with type 2 diabetes using SGLT2is and strong statins than in those using metformin and strong statins. Thus, we would like to emphasize that clinicians should pay attention to patients' foot conditions in routine clinical practice when they prescribe SGLT2is, particularly in female patients with type 2 diabetes and patients with type 2 diabetes with suspected advanced atherosclerosis requiring strong statins.
Conversely, we could not clarify the risk of lower‐limb amputation associated with SGLT2is in patients with type 2 diabetes at high risk for cardiorenal complications in the present study due to a lack of clinical data, such as duration of diabetes, smoking status, glycated hemoglobin levels and micro‐ or macrovascular complications, which is a limitation of the present study, using DPC‐based clinical databases. However, considering that the use of SGLT2is is currently essential to patients with type 2 diabetes to prevent cardiorenal complications, who are generally considered at high risk for micro‐/macrovascular complications including PAD, clinicians should consider the risk of lower‐limb amputation when using SGLT2is in patients with type 2 diabetes.
DISCLOSURE
AS received lecture fees from Astellas Pharma Inc., Eli Lilly Japan K.K., Novo Nordisk Pharma Inc. and Sanofi K.K. TH received lecture fees from Kowa K.K. The other authors declare no conflict of interest.
Approval of the research protocol: The ethics board of Kitasato University admitted that the protocol for this study does not require ethics approval, because all available data are completely anonymous with no personal information, which is characteristic of DPC‐based clinical databases (Control number: B19‐285, dated 31 January 2020). This study was carried out in accordance with the Declaration of Helsinki, and the ethical guidelines for medical and health research involving human subjects.
Informed consent: All patient data were anonymized and contained no personal data; thus, informed consent was not required.
Registry and the registration no. of the study/trial: N/A.
Animal studies: N/A.
Supporting information
Table S1 | Baseline clinical characteristics of patients with type 2 diabetes using metformin and sodium–glucose cotransporter 2 inhibitors before 1:1 propensity score matching, and the number of those who underwent lower‐limb amputation during the observation period.
ACKNOWLEDGMENT
No specific funding or grants were received for the study.
REFERENCES
- 1. Li D, Yang JY, Wang T, et al. Risks of diabetic foot syndrome and amputation associated with sodium glucose co‐transporter 2 inhibitors: a Meta‐analysis of Randomized Controlled Trials. Diabetes Metab 2018; 44: 410–414. [DOI] [PubMed] [Google Scholar]
- 2. Khouri C, Cracowski JL, Roustit M. SGLT‐2 inhibitors and the risk of lower‐limb amputation: is this a class effect? Diabetes Obes Metab 2018; 20: 1531–1534. [DOI] [PubMed] [Google Scholar]
- 3. Higashi Y, Miyata T, Shigematsu H, et al. Evaluation of risk factors for major amputation in patients with diabetes and peripheral artery disease receiving antiplatelet therapy‐ post hoc analysis of a prospective observational multicenter cohort Study (SEASON). Circ J 2019; 83: 1929–1936. [DOI] [PubMed] [Google Scholar]
- 4. van Bommel EJ, Muskiet MH, Tonneijck L, et al. SGLT2 inhibition in the diabetic kidney‐from mechanisms to clinical outcome. Clin J Am Soc Nephrol 2017; 12: 700–710. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Saisho Y. SGLT2 inhibitors: the star in the treatment of type 2 diabetes? Diseases 2020; 8: 14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Neal B, Perkovic V, Mahaffey KW, et al. Canagliflozin and cardiovascular and renal events in type 2 diabetes. N Engl J Med 2017; 377: 644–657. [DOI] [PubMed] [Google Scholar]
- 7. Chang HY, Singh S, Mansour O, et al. Association between sodium‐glucose cotransporter 2 inhibitors and lower extremity amputation among patients with type 2 diabetes. JAMA Intern Med 2018; 178: 1190–1198. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Chatterjee S, Bandyopadhyay D, Ghosh RK, et al. SGLT‐2 inhibitors and peripheral artery disease: a statistical hoax or reality? Curr Probl Cardiol 2019; 44: 207–222. [DOI] [PubMed] [Google Scholar]
- 9. Miyashita S, Kuno T, Takagi H, et al. Risk of amputation associated with sodium‐glucose co‐transporter 2 inhibitors: a meta‐analysis of five randomized controlled trials. Diabetes Res Clin Pract 2020; 163: 108136. [DOI] [PubMed] [Google Scholar]
- 10. Heyward J, Mansour O, Olson L, et al. Association between sodium‐glucose cotransporter 2 (SGLT2) inhibitors and lower extremity amputation: a systematic review and meta‐analysis. PLoS One 2020; 15: e0234065. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Lin C, Zhu X, Cai X, et al. SGLT2 inhibitors and lower limb complications: an updated meta‐analysis. Cardiovasc Diabetol 2021; 20: 91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Examination Committee of Criteria for 'Obesity Disease' in Japan, Japan Society for the Study of Obesity . New criteria for 'obesity disease' in Japan. Circ J 2002; 66: 987–992. [DOI] [PubMed] [Google Scholar]
- 13. Austin PC. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav Res 2011; 46: 399–424. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Hippisley‐Cox J, Coupland C. Diabetes treatments and risk of amputation, blindness, severe kidney failure, hyperglycaemia, and hypoglycaemia: open cohort study in primary care. BMJ 2016; 352: i1450. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Radholm K, Zhou Z, Clemens K, et al. Effects of sodium‐glucose co‐transporter‐2 inhibitors in type 2 diabetes in women versus men. Diabetes Obes Metab 2020; 22: 263–266. [DOI] [PubMed] [Google Scholar]
- 16. Erkens JA, Klungel OH, Stolk RP, et al. Antihypertensive drug therapy and the risk of lower extremity amputations in pharmacologically treated type 2 diabetes patients. Pharmacoepidemiol Drug Saf 2004; 13: 139–146. [DOI] [PubMed] [Google Scholar]
- 17. Tani Y, Asakura K, Sasaki S, et al. The influence of season and air temperature on water intake by food groups in a sample of free‐living Japanese adults. Eur J Clin Nutr 2015; 69: 907–913. [DOI] [PubMed] [Google Scholar]
- 18. Ohno K, Asami M, Matsui Y. Is the default of 2 liters for daily per‐capita water consumption appropriate? A nationwide survey reveals water intake in Japan. J Water Health 2018; 16: 562–573. [DOI] [PubMed] [Google Scholar]
- 19. Moss SE, Klein R, Klein BE. The 14‐year incidence of lower‐extremity amputations in a diabetic population. The Wisconsin Epidemiologic Study of Diabetic Retinopathy. Diabetes Care 1999; 22: 951–959. [DOI] [PubMed] [Google Scholar]
- 20. Misher A, Ampuero A, Phan R, et al. Sodium‐glucose cotransporter‐2 inhibitors and the risk of amputation: what is currently known? Am J Ther 2020; 28: e96–e110. [DOI] [PubMed] [Google Scholar]
- 21. Lin C, Liu J, Sun H. Risk factors for lower extremity amputation in patients with diabetic foot ulcers: a meta‐analysis. PLoS One 2020; 15: e0239236. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Lin CW, Hsu BR, Tsai JS, et al. Effect of limb preservation status and body mass index on the survival of patients with limb‐threatening diabetic foot ulcers. J Diabetes Complications 2017; 31: 180–185. [DOI] [PubMed] [Google Scholar]
- 23. Sohn MW, Budiman‐Mak E, Lee TA, et al. Significant J‐shaped association between body mass index (BMI) and diabetic foot ulcers. Diabetes Metab Res Rev 2011; 27: 402–409. [DOI] [PubMed] [Google Scholar]
- 24. Kashiwagi A, Yoshida S, Nakamura I, et al. Efficacy and safety of ipragliflozin in Japanese patients with type 2 diabetes stratified by body mass index: a subgroup analysis of five randomized clinical trials. J Diabetes Investig 2016; 7: 544–554. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Ministry of Health, Labour and Welfare . National Health and Nutrition Survey 2017. Available from: https://www.mhlw.go.jp/bunya/kenkou/kenkou_eiyou_chousa.html Accessed April 1, 2022.
- 26. Ying AF, Tang TY, Jin A, et al. Diabetes and other vascular risk factors in association with the risk of lower extremity amputation in chronic limb‐threatening ischemia: a prospective cohort study. Cardiovasc Diabetol 2022; 21: 7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Beckman JA, Duncan MS, Damrauer SM, et al. Microvascular disease, peripheral artery disease, and amputation. Circulation 2019; 140: 449–458. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Marso SP, Hiatt WR. Peripheral arterial disease in patients with diabetes. J Am Coll Cardiol 2006; 47: 921–929. [DOI] [PubMed] [Google Scholar]
- 29. Yen FS, Wei JC, Shih YH, et al. The risk of nephropathy, retinopathy, and leg amputation in patients with diabetes and hypertension: a nationwide, population‐based retrospective cohort study. Front Endocrinol (Lausanne) 2021; 12: 756189. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Adler AI, Boyko EJ, Ahroni JH, et al. Lower‐extremity amputation in diabetes. The independent effects of peripheral vascular disease, sensory neuropathy, and foot ulcers. Diabetes Care 1999; 22: 1029–1035. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1 | Baseline clinical characteristics of patients with type 2 diabetes using metformin and sodium–glucose cotransporter 2 inhibitors before 1:1 propensity score matching, and the number of those who underwent lower‐limb amputation during the observation period.