Abstract
Background
Several previous studies have identified a potential risk of acute kidney injury (AKI) associated with sodium‐glucose cotransporter‐2 (SGLT‐2) inhibitors, based on adverse event reports. However, recent European observational studies have shown conflicting results.
Objective
To evaluate the risk of AKI in patients with type 2 diabetes (T2DM) who were treated with dapagliflozin compared with sitagliptin.
Method
We conducted a retrospective cohort study on patients with T2DM who were newly prescribed dapagliflozin or sitagliptin between September 1, 2014, and June 30, 2021, using the nationwide National Health Insurance Review and Assessment (HIRA) Service database in Korea. Propensity scores were estimated using a multivariable logistic regression model, and matching was performed at a 1:1 ratio to balance the dapagliflozin and sitagliptin groups. The outcome of interest was the occurrence of AKI hospitalization 90 days post‐exposure, captured by a validated algorithm based on the International Classification of Diseases 10th Revision (ICD‐10) code: N17. Hazard ratios (HR) with 95% confidence intervals (CI) were calculated using a Cox proportional hazards model.
Results
Among 94,977 dapagliflozin users matched to sitagliptin users, AKI events occurred in 132 dapagliflozin users versus 198 sitagliptin users, with incidence rates of 2.92 and 8.93 per 1000 person‐years, respectively. The risk of AKI events was 34% lower in dapagliflozin users (HR: 0.66, 95% CI: 0.53–0.83) compared with sitagliptin users. This protective effect remained consistent in sensitivity analyses.
Conclusion
Contrary to the United States Food and Drug Administration's safety warning, our findings suggest that dapagliflozin may have a protective effect against AKI in patients with T2DM. This is consistent with recent findings from European post‐marketing safety studies and may serve as supportive evidence.
Keywords: acute kidney injury, cohort study, dipeptidyl peptidase 4 inhibitor, sodium‐glucose transporter 2 inhibitors
1. BACKGROUND
Dapagliflozin, a sodium‐glucose cotransporter‐2 (SGLT‐2) inhibitor used to treat type 2 diabetes mellitus (T2DM), lowers blood glucose levels by blocking the reabsorption of glucose in the kidneys, prompting its excretion in the urine. 1 This drug was effective in lowering blood glucose levels, but several safety issues were identified during clinical trials. When the European Medicines Agency (EMA) first approved dapagliflozin in November 2012, they required additional non‐interventional post‐authorization safety studies (PASS) to address unresolved safety concerns from clinical trials, including urinary tract infections, acute kidney injury (AKI), acute liver injury, and bladder cancer. 2
AKI is characterized by a sudden decline in kidney function, resulting in the accumulation of waste products and potential electrolyte imbalances in the body. Although SGLT‐2 inhibitors, including dapagliflozin, have been effective in managing blood glucose levels, safety concerns about AKI have been raised because of its biological mechanism of action on the kidneys. By blocking the reabsorption of glucose in the kidney, elevated glucose concentration in nephron segments can lead to osmotic diuresis, natriuresis, and volume depletion, ultimately leading to renal vasoconstriction and decreased renal perfusion. 3 , 4 The reduction in renal perfusion pressure can potentially contribute to renal ischemia causing AKI, particularly in patients with preexisting renal impairment or other risk factors for AKI. 5 , 6 This concern was further strengthened by both a meta‐analysis conducted in 2013 and a warning issued based on the United States Food and Drug Administration (FDA) adverse effect reporting system (FAERS) in June 2016. 7 , 8 According to a disproportionality analysis of acute renal failure (ARF) cases submitted to the FAERS between January 2013 and September 2016, cases involving SGLT‐2 inhibitors showed a significantly higher proportion of ARF compared to cases without SGLT2 inhibitors in patients with T2DM. 9 In contrast, the recently completed EMA PASS reported that dapagliflozin reduced the risk of AKI by 30% compared to other glucose‐lowering drugs and by 31% compared to the placebo group. 10 , 11
In Korea, dapagliflozin was first approved on November 26, 2013; post‐marketing surveillance was conducted to identify potential adverse events among 3371 patients who started treatment with dapagliflozin over 6 years of post‐approval, and several cases of acute pyelonephritis were identified from this surveillance. 12 , 13 However, the absence of a comparison group in Korean case reports made it difficult to conclude the association between AKI and dapagliflozin. In addition, several observational studies conducted in Asia have shown conflicting results. 14 , 15 Therefore, this study aimed to investigate the association between dapagliflozin use and the risk of AKI in patients with T2DM using large‐scale, nationwide claims data from Korea.
2. METHODS
2.1. Data source
This study utilized the nationwide health insurance claims database of the Health Insurance Review and Assessment (HIRA) from September 1, 2013, to December 31, 2021. HIRA claims data are collected when health‐care providers in Korea request reimbursement for health‐care services covered by the National Health Insurance Service (NHIS). Korea operates a single‐payer public health insurance system, with approximately 97% of the Korean population enrolled in the NHIS. Under this system, insured citizens pay only 5%–30% of the health‐care services they receive, and health‐care providers submit claims to HIRA for reimbursement of the remaining health‐care costs. HIRA reviews these claims to ensure the quality, appropriateness, and accuracy of the medical services provided. Information related to patient demographics, disease details, treatment details, surgical procedures, and more are accumulated in the HIRA database throughout this process. 16 These data are de‐identified to protect patient privacy and provided to researchers for research purposes. Patient consent was not required for analyses of secondary claims data from de‐identified human records, and we received confirmation of research ethics exemption from the Institutional Review Board of Ewha Womans University (IRB number ewha‐202204‐0008‐01).
2.2. Study design and population
This study utilized a retrospective cohort study design, entering patients aged over 18 years who initiated a single component of either dapagliflozin or sitagliptin between September 1, 2014 (the date of Korean insurance coverage initiation for dapagliflozin) and June 30, 2021. The start date of treatment with either dapagliflozin or sitagliptin was set as the index date for each patient. Patients were excluded if they had ever taken an SGLT‐2 inhibitor, including dapagliflozin, or a dipeptidyl peptidase 4 (DPP‐4) inhibitor, including sitagliptin, before the index date. Those without a diagnosis code of T2DM (International Classification of Diseases 10th Revision [ICD‐10] code: E11) or those diagnosed with type 1 diabetes (ICD‐10 code: E10) before the index date were also excluded. Finally, patients who had a prescription for both dapagliflozin and sitagliptin concurrently on the index date, or those with a history of AKI within 1 year before or on the index date, were excluded from the study population. The study design with the inclusion and exclusion criteria for selecting eligible study participants is shown in Figure S1. The dapagliflozin and sitagliptin treatment groups were matched in a 1:1 ratio based on propensity scores (PS) of covariates to achieve a balance between the two groups.
2.3. Exposures and outcomes
The exposure drug was single‐component dapagliflozin, and the comparator drug was single‐component sitagliptin. Fixed‐dose combination drugs containing dapagliflozin or sitagliptin, such as those combined with metformin, were excluded from both the exposure and comparator drugs. Other glucose‐lowering drugs, if taken separately and not as fixed‐dose combinations, were also excluded. Sitagliptin is a drug that lowers blood sugar in patients with T2DM by inhibiting DPP‐4, an enzyme that degrades glucagon‐like peptide‐1 (GLP‐1), a hormone that stimulates insulin secretion to lower blood sugar levels. Sitagliptin has been used for a long time since its first approval in Korea in September 2007 and has been demonstrated by international studies not to be associated with AKI. 17 , 18 In addition, it was selected as the comparator drug for dapagliflozin because it was the most commonly used second‐line treatment for T2DM in Korea during our study period. The outcome of interest was the occurrence of AKI, defined as a first hospitalization with ICD‐10 code ‘N17’ as a primary or secondary diagnosis. 19 , 20 All patients were followed from the index date until the earliest of the following events: AKI onset, 90 days of post‐exposure, the end of the database period (December 31, 2021), or a drug switch from dapagliflozin to DPP‐4 inhibitors or from sitagliptin to SGLT‐2 inhibitors. Our primary exposure assessment window was 90 days from the index date, and we varied the window to the following periods in sensitivity analyses: 30, 180, 270, and 365 days.
2.4. Covariates
Baseline characteristics measured on or before the index date of our study population, including age, sex, health insurance status, type of medical institution, Charlson Comorbidity Index (CCI), and co‐medication status, were selected as covariates. Comorbidity conditions using the CCI were assessed during the 365 days before the index date. 21 Co‐medication status was assessed within 365 days before the index date for medications known to be associated with AKI, as well as other anti‐diabetic drugs, excluding SGLT‐2 inhibitors and DPP‐4 inhibitors. The assessed medications included beta‐blockers, bisphosphonates, calcium channel blockers, digoxin, nitrates, anticonvulsants, nonsteroidal anti‐inflammatory drugs (NSAIDs), antidepressants, immunosuppressants, antiarrhythmic agents, diuretics, angiotensin receptor blocker (ARB), angiotensin‐converting enzyme inhibitor (ACEI), antiplatelet agents, anti‐ulcer agents, anticoagulants, agents for gout, antineoplastic agents, antibiotics, antifungals, and other anti‐diabetic drugs (biguanides, sulfonylureas, meglitinides, thiazolidinedione [TZDs], alpha‐glucosidase inhibitors, GLP‐1 agonists, insulin). 22 , 23 , 24
2.5. Statistical analysis
Descriptive analyses were conducted to compare the baseline characteristics of the dapagliflozin and sitagliptin groups. For categorical variables, a chi‐square test was performed to analyze differences in distributions, and for continuous variables, a t‐test was used. The balance between the two groups was assessed using overlaid histograms and absolute standardized differences (aSD), considering values <0.1 as indicative of proper balance. To minimize potential confounding factors, PS for each group was computed using multivariate logistic regression models with all baseline characteristics as independent variables and treatment drugs including dapagliflozin and sitagliptin as the dependent variables. The calculated PS were then matched using the greedy matching method in a 1:1 ratio. The Cox proportional hazards model was used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for balanced groups regarding the risk of AKI occurrence. Two‐sided p‐values < 0.05 were considered statistically significant. The proportional hazards assumption was verified using a Cox proportional hazards model that included an interaction term between the treatment drug and the logarithm of time to check for normality. If the interaction term was not significant (p‐value > 0.05), it supported the proportional hazards assumption, indicating that the treatment effect was consistent over time. Furthermore, sensitivity analyses were conducted to test the robustness of our findings by modifying the follow‐up period from 90 to 30 days, 180, 270, and 365 days. Subgroup analyses were performed based on characteristics known as major risk factors for AKI, including age, insulin use, presence of chronic kidney disease, gender, and CCI. All analyses were performed using SAS 9.4 (SAS Institute, Cary, North Carolina) and visualization using R studio (version 2023.06.1). This article follows the Strengthening the Reporting of Observational studies in Epidemiology (STROBE) guidelines. 25
3. RESULTS
3.1. Basic characteristics of the study population
We identified 107,512 patients with T2DM who initiated dapagliflozin therapy and 182,466 patients who initiated sitagliptin therapy between September 1, 2014, and June 30, 2021. Before PS matching, dapagliflozin users were younger, had lower CCI, and were more likely to have experienced calcium channel blockers and antiplatelet agents use within 1 year prior to treatment initiation compared to sitagliptin users. The proportion of beta‐blockers, bisphosphonates, anticonvulsants, antidepressants, diuretics, antineoplastic agents, sulfonylurea, and insulin use was higher in the sitagliptin group. The gender distribution was similar between the two groups, with males comprising a higher proportion than females in both groups. After PS matching, 94,977 dapagliflozin users were matched 1:1 to sitagliptin users (Figure 1). The baseline characteristics between the two groups were well balanced, with absolute standardized differences <0.1 for all characteristics (Table 1) (Figure S2).
FIGURE 1.
Flow chart of the study population. Study populations were selected and matched 1:1 based on the propensity score calculated by the logistic regression including covariates. AKI, acute kidney injury; DPP‐4, dipeptidyl peptidase‐4; SGLT‐2, sodium‐glucose cotransporter‐2; T2DM, type 2 diabetes mellitus.
TABLE 1.
Basic characteristics of study population at index date or baseline (dapagliflozin vs. sitagliptin).
| Before PS matching | After PS matching | |||||
|---|---|---|---|---|---|---|
| Dapagliflozin (n = 107,512) | Sitagliptin (n = 182,466) | SD | Dapagliflozin (n = 94,977) | Sitagliptin (n = 94,977) | aSD | |
| N (%) | N (%) | N (%) | N (%) | |||
| Age, years | ||||||
| Mean (std) | 54.3 (13.3) | 61.9 (13.3) | −0.570 | 56.3 (12.4) | 56.3 (12.8) | 0.002 |
| <45 | 24,498 (22.8) | 18,002 (9.9) | −0.469 | 15,976 (16.8) | 16,411 (17.3) | −0.018 |
| 45–54 | 28,415 (26.4) | 34,375 (18.8) | 25,166 (26.5) | 25,589 (26.9) | ||
| 55–64 | 30,982 (28.8) | 50,966 (27.9) | 30,256 (31.9) | 28,677 (30.2) | ||
| ≥65 | 23,617 (22.0) | 79,123 (43.4) | 23,579 (24.8) | 24,300 (25.6) | ||
| Gender | ||||||
| Male | 60,595 (56.4) | 102,314 (56.3) | 0.006 | 53,659 (56.5) | 54,175 (57.0) | −0.011 |
| CCI | ||||||
| Mean (std) | 0.93 (1.29) | 1.24 (1.68) | −0.212 | 1.0 (1.3) | 1.0 (1.4) | −0.001 |
| ≥5 | 2135 (2.0) | 8243 (4.5) | −0.143 | 2072 (2.2) | 2486 (2.6) | −0.029 |
| Health insurance status | ||||||
| Medical aid | 5391 (5.0) | 12,427 (6.8) | −0.076 | 5039 (5.3) | 5092 (5.4) | −0.003 |
| Medical institution | ||||||
| Tertiary hospital | 47,858 (44.5) | 71,861 (39.4) | −0.084 | 40,064 (42.2) | 39,688 (41.8) | 0.000 |
| Secondary hospital | 7131 (6.6) | 24,698 (13.5) | 7027 (7.4) | 6995 (7.4) | ||
| Clinic | 51,596 (48.0) | 82,540 (45.2) | 46,959 (49.4) | 47,363 (49.9) | ||
| Public health care | 937 (0.9) | 3367 (1.9) | 927 (1.0) | 931 (1.0) | ||
| Medication status | ||||||
| Beta‐blocker | 62,564 (58.2) | 112,005 (61.4) | −0.065 | 55,396 (58.3) | 55,294 (58.2) | 0.002 |
| Bisphosphonates | 2442 (2.3) | 8487 (4.7) | −0.131 | 2434 (2.6) | 2396 (2.5) | 0.003 |
| Calcium channel blockers | 40,116 (40.7) | 74,241 (37.3) | −0.069 | 35,703 (37.6) | 35,609 (37.5) | 0.002 |
| Digoxin | 3226 (1.3) | 1386 (1.8) | −0.039 | 1290 (1.4) | 1291 (1.4) | 0.000 |
| Nitrates | 6822 (6.4) | 9315 (5.1) | 0.053 | 5677 (6.0) | 5682 (6.0) | 0.000 |
| Anticonvulsants | 9003 (8.4) | 21,880 (12.0) | −0.120 | 8425 (8.9) | 8500 (9.0) | −0.003 |
| NSAIDs | 89,321 (83.1) | 155,371 (85.2) | −0.057 | 79,229 (83.4) | 79,024 (83.2) | 0.006 |
| Antidepressants | 35,183 (32.7) | 73,152 (40.1) | −0.154 | 32,247 (34.0) | 32,125 (33.8) | 0.003 |
| Immunosuppressants | 1069 (1.0) | 2788 (1.5) | −0.048 | 1020 (1.1) | 1060 (1.1) | −0.004 |
| Antiarrhythmic agents | 6026 (5.6) | 12,460 (6.8) | −0.051 | 5459 (5.8) | 5466 (5.8) | 0.000 |
| Diuretics | 15,076 (14.0) | 33,657 (18.5) | −0.120 | 13,773 (14.5) | 13,818 (14.6) | −0.001 |
| ARB/ACEI | 58,406 (54.3) | 98,301 (53.9) | 0.009 | 51,413 (54.1) | 51,270 (54.0) | 0.003 |
| Antiplatelet agents | 73,068 (68.0) | 111,204 (61.0) | 0.147 | 63,616 (67.0) | 63,669 (67.0) | −0.001 |
| Anti‐ulcer agents | 86,865 (80.8) | 152,541 (83.6) | −0.073 | 77,313 (81.4) | 77,141 (81.2) | 0.005 |
| Anticoagulants | 7819 (7.3) | 16,695 (9.2) | −0.068 | 7035 (7.4) | 7061 (7.4) | −0.001 |
| Agents for gout | 3498 (3.3) | 5606 (3.1) | 0.010 | 3005 (3.2) | 3021 (3.2) | −0.001 |
| Antineoplastic agents | 1658 (1.5) | 7382 (4.1) | −0.152 | 1636 (1.7) | 1701 (1.8) | −0.005 |
| Antibiotics | 69,178 (64.3) | 122,108 (66.9) | −0.054 | 61,243 (64.5) | 61,094 (64.3) | 0.003 |
| Antifungal | 32,108 (29.9) | 61,240 (33.6) | −0.080 | 28,795 (30.3) | 28,791 (30.3) | 0.000 |
| Other Ads | ||||||
| Biguanide | 79,252 (73.7) | 134,864 (73.9) | −0.005 | 70,084 (73.8) | 69,922 (73.6) | 0.004 |
| Sulfonylurea | 35,622 (33.1) | 78,477 (43.0) | −0.204 | 33,598 (35.4) | 34,285 (36.1) | −0.015 |
| Meglitinide | 639 (0.6) | 1585 (0.9) | −0.031 | 624 (0.7) | 625 (0.7) | 0.000 |
| TZD | 8210 (7.6) | 12,219 (6.7) | 0.036 | 7193 (7.6) | 7319 (7.7) | −0.005 |
| Alpha‐glucosidase inhibitor | 3948 (3.7) | 8273 (4.5) | −0.044 | 3741 (3.9) | 3780 (4.0) | −0.002 |
| GLP‐1 agonist | 445 (0.4) | 139 (0.1) | 0.068 | 138 (0.2) | 133 (0.1) | 0.001 |
| Insulin | 16,121 (15.0) | 41,546 (22.8) | −0.200 | 15,041 (15.8) | 15,084 (15.9) | −0.001 |
Abbreviations: ACEI, angiotensin‐converting enzyme inhibitor; Ads, anti‐diabetic agents; ARB, angiotensin receptor blocker; aSD, absolute standard difference; CCI, Charlson comorbidity index; GLP‐1, glucagon‐like peptide‐1; NSAIDS, nonsteroidal anti‐inflammatory drug; PS, Propensity score; SD, standard difference; std., Standard deviation; TZD, thiazolidinedione.
3.2. Association between dapagliflozin treatment and hospitalized AKI events
In primary analysis among the matched study population, 132 cases of hospitalized AKI occurred within 90 days after the initiation of dapagliflozin treatment, and 198 hospitalized AKI cases occurred in the sitagliptin group. The incidence rates per 1000 person‐years (PY) were 5.92 and 22.18, respectively. Because the proportional hazards assumption was fully satisfied (p = 0.862), we performed the analysis to estimate the HRs for AKI hospitalization using the Cox proportional hazard model, showing that the dapagliflozin group had a 34% lower risk of hospitalization for AKI within 90 days compared to the sitagliptin group (HR 0.66; 95% CI 0.53–0.83) (Figure 2). In sensitivity analyses, where the follow‐up windows were varied to 30, 180, 270, and 365 days, the protective effects of dapagliflozin were consistently maintained. In subgroup analyses, the primary outcome of dapagliflozin's effect on AKI showed inconsistency across different strata including age, gender, CCI, history of chronic kidney disease (CKD)CC, and previous insulin use. Although individuals aged 65 years or older, females, those with a CCI of 5 or higher, and those with previous insulin use showed HR estimates below the null, their confidence intervals spanned the null. In patients with preexisting CKD before initiating dapagliflozin, no protective effect against AKI was observed (HR 1.11, 95% CI 0.58–2.11) (Figure 3). However, a consistent protective effect against AKI was noted in both users and non‐users of metformin, ARBs/ACEIs, and diuretics.
FIGURE 2.
Risk of AKI with dapagliflozin compared to sitagliptin. Risks of AKI occurrence with dapagliflozin compared to sitagliptin were calculated according to the follow‐up period after the index date using a Cox proportional hazard model. The incidence rate per 1000 person‐years was estimated by dividing the number of AKI cases by the total person‐years and multiplying by 1000 in each cohort. CI, confidence interval; HR, hazard ratio; IR, incidence ratio; PY, person year.
FIGURE 3.
Risk of AKI with dapagliflozin compared to sitagliptin stratified by baseline characteristics. The risk of AKI with dapagliflozin and sitagliptin was calculated using the Cox proportional hazards model after stratifying by baseline characteristics. The incidence rate per 1000 person‐years was estimated by dividing the number of AKI cases by the total person‐years and multiplying by 1000 in each cohort. CCI was identified for each patient until 365 days before the index date, including the index date. CKD prevalence, insulin use, ARB/ACEI use, diuretic use, and metformin use were identified for each patient until 365 days before the index date, excluding the index date. ACEI, angiotensin‐converting enzyme inhibitor; ARB, angiotensin receptor blocker; CCI, Charlson comorbidity index; CI, confidence interval; CKD, chronic kidney disease; HR, hazard ratio; PY, person year.
4. DISCUSSION
We conducted a retrospective cohort study using nationwide claims data to assess whether the risk of AKI differed between patients with T2DM who initiated dapagliflozin or sitagliptin between September 1, 2014, and June 30, 2021. In a cohort of 94,977 dapagliflozin users and their 1:1 matched sitagliptin comparators, our results showed a 34% reduction in the risk of AKI hospitalization within 90 days of dapagliflozin initiation, with this reduction remaining consistent across various follow‐up periods.
Most previous trials and studies have reported that SGLT‐2 inhibitors do not increase the risk of AKI. In a large clinical trial, the Dapagliflozin Effect on Cardiovascular Events–Thrombolysis in Myocardial Infarction 58 (DECLARE‐TIMI 58), dapagliflozin use was associated with a lower incidence of AKI events compared to placebo. 11 Early studies conducted for 1207 and 4778 SGLT‐2 inhibitor users in the United States and Canada, respectively, using data from electronic medical records found a trend toward a reduced risk of AKI in SGLT‐2 inhibitor users compared to non‐users or other oral glucose‐lowering drugs (oGLDs) users, though the results were not statistically significant. 26 , 27 Recent large‐scale observational studies across diverse racial groups have reported a protective effect of dapagliflozin against AKI. In Europe, a PASS analyzing data from approximately 34,000 dapagliflozin users found a 30% reduction in AKI risk compared to all oGLDs and a 22% reduction compared to DPP‐4 inhibitors. 10 A Taiwanese study including a total of 29,116 dapagliflozin‐treated patients derived from a national claims database also demonstrated a 39% reduction in AKI risk with dapagliflozin compared to DPP‐4 inhibitors. 15 Our findings align with these results in terms of trends and the magnitude of estimates, further reinforcing evidence of the protective effects of SGLT‐2 inhibitors.
Older age and male gender are well‐known risk factors for AKI. 28 , 29 , 30 A previous Japanese study found that characteristics such as being over 65 years of age, male, having comorbidities, and low estimated glomerular filtration rate (eGFR) levels contribute to a higher incidence of AKI among users of SGLT‐2 inhibitors. 14 However, the Japanese study did not include a comparison group for SGLT‐2 inhibitors and did not exclude a history of AKI prior to taking SGLT‐2 inhibitors, making it difficult to determine whether SGLT‐2 inhibitors truly contributed to AKI incidence. According to our research, dapagliflozin contributes to the reduction of AKI, with a particularly strong protective effect observed in men. The mechanism behind the gender‐specific effects of dapagliflozin remains unknown, so further research is needed. Additionally, although the Japanese study reported a significant increase in AKI risk when SGLT‐2 inhibitors were used in combination with diuretics, our study results showed no difference in AKI risk based on diuretic use.
In our study, the only patient group with an HR point estimate for AKI exceeding 1 was the group with CKD. The Dapagliflozin And Prevention of Adverse outcomes in Chronic Kidney Disease (DAPA‐CKD) randomized clinical trials, which evaluated the efficacy and safety of canagliflozin and dapagliflozin in patients with CKD, reported initial mean eGFR reductions of 3.7 mL/min/1.73 m2 within 3 weeks and 4.0 mL/min/1.73 m2 within 2 weeks after dapagliflozin treatment, respectively. However, importantly, in patients with CKD, the incidence of AKI was significantly lower in the dapagliflozin‐treatment group compared to the placebo group (HR 0.68, 95% CI 0.49–0.94). 31 , 32 , 33 Likewise, systematic meta‐analyses have indicated that SGLT‐2 inhibitors provide a protective effect against AKI in patients with CKD compared to placebo (Relative risk, RR 0.82, 95% CI 0.72–0.93, p = 0.003). 34 , 35 , 36 Contrary to this evidence, our result did not show a protective effect with dapagliflozin in patients with CKD, with a wide confidence interval (HR 1.11, 95% CI 0.58–2.11). Overall, these results are important for clinical practice, suggesting that dapagliflozin can be safely used in patients with CKD without increasing the risk of AKI. Therefore, the use of SGLT‐2 inhibitors in patients with CKD should not be discontinued or rejected.
Our study has several limitations that should be considered when interpreting the findings. First, because the claims database lacks laboratory data and information on disease staging or severity, AKI occurrence was defined using ICD‐10 codes. To minimize misclassification, this study identified AKI cases based on hospital admissions where AKI was recorded as either a primary or secondary diagnosis. Instead, this restriction may have led to an underestimation of milder AKI cases that did not require clinical treatment. If the outcome definition is highly specific, bias is likely to be minimal, and estimates are toward the null. Despite this, the current study demonstrated a protective effect of dapagliflozin compared to sitagliptin. Second, the exposure group was assigned based on dapagliflozin prescription status on the index date, without accounting for the prescription duration, adherence, or dosage. Although prescription duration and adherence were not considered, their impact is likely minimal given the relatively short follow‐up period. Moreover, sensitivity analyses with both shorter and longer follow‐up durations consistently yielded similar results. Most patients are likely to start with the low dose, and the impact of dose titration on the outcome might be minimal given the short follow‐up period. However, as little is known about the association between dapagliflozin dosage and AKI risk, further research may be warranted. Finally, despite using PS matching to balance observed confounders between treatment groups, residual confounding remains a possibility. Variables such as baseline kidney function, body mass index, and lifestyle factors (e.g., smoking and physical activity) were not available in the claims data and may have influenced the observed associations. However, both SGLT‐2 inhibitors and DPP‐4 inhibitors are primarily second‐line therapies in Korea and are more expensive than metformin or sulfonylureas. As a result, these drugs are likely to be prescribed to patients at similar stages of T2DM. Selecting appropriate comparison groups, combined with a new‐user design and 1:1 PS matching on multiple proxies of diabetes severity and duration, has been shown to significantly reduce bias. 37 Accordingly, our methods minimize concerns regarding unmeasured confounding.
Nevertheless, the strength of our study lies in the use of a large research population, including approximately 95,000 dapagliflozin users in Korea, which enhances the robustness and generalizability of our findings. This large sample size allowed for subgroup analyses based on patient characteristics, revealing potential variations in the effects on AKI according to age, gender, comorbidities, insulin use, and CKD history.
5. CONCLUSIONS
Compared to sitagliptin, dapagliflozin was associated with a lower risk of AKI‐related hospitalization in patients with T2DM. This finding is consistent with multinational studies from Taiwan, the United States, Europe, and Canada, suggesting generalizability across different ethnicities and populations. However, stratified analyses indicated potential variations in effects based on baseline factors such as age, sex, and comorbidities. Although the confidence intervals were wide, the observed trends highlight the need for further research to confirm these potential differences.
AUTHOR CONTRIBUTIONS
All authors attest they meet the ICMJE criteria for authorship. Conceptualization: Hee‐Jin Kim, HoJin Shin, Ye‐Jee Kim, Nam‐Kyong Choi; Investigation: Heehyun Won, Suvin Park, Hui‐Eon Lee, Haerin Cho, Jeong Ah Kim; Methodology: Hee‐Jin Kim, Hui‐Eon Lee, HoJin Shin, Ye‐Jee Kim, Nam‐Kyong Choi; Data curation: Hee‐Jin Kim; Formal analysis: Hee‐Jin Kim; Writing‐original draft: Hee‐Jin Kim; Writing‐review & editing: Hee‐Jin Kim, Na‐Young Jeong, HoJin Shin, Ye‐Jee Kim; Supervision: Nam‐Kyong Choi.
FUNDING INFORMATION
This research was funded by the Ministry of Food and Drug Safety (grant number: 21183MFDS543) in the Republic of Korea.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest.
Supporting information
Figure S1.–S2.
ACKNOWLEDGMENTS
The authors thank the Ministry of Food and Drug Safety for funding and the National Health Insurance Review and Assessment for data source provision in the Republic of Korea.
Kim H‐J, Won H, Park S, et al. Risk of acute kidney injury in dapagliflozin users with type 2 diabetes: A nationwide propensity score‐matched cohort study in Korea. Pharmacotherapy. 2025;45:282‐290. doi: 10.1002/phar.70015
DATA AVAILABILITY STATEMENT
The database is not available due to the privacy policy.
REFERENCES
- 1. Xu B, Li S, Kang B, Zhou J. The current role of sodium‐glucose cotransporter 2 inhibitors in type 2 diabetes mellitus management. Cardiovasc Diabetol. 2022;21(1):83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. European Medicines Agency . Forxiga. 2012. Accessed June 1, 2024. https://www.ema.europa.eu/en/documents/assessment‐report/forxiga‐epar‐public‐assessment‐report_en.pdf
- 3. Lytvyn Y, Bjornstad P, van Raalte DH, Heerspink HL, Cherney DZ. The new biology of diabetic kidney disease—mechanisms and therapeutic implications. Endocr Rev. 2020;41(2):202‐231. doi: 10.1210/endrev/bnz010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Bommel E, Lytvyn Y, Perkins BA. Renal hemodynamic effects of sodium‐glucose cotransporter 2 inhibitors in hyperfiltering people with type 1 diabetes and people with type 2 diabetes and normal kidney function. Kidney Int. 2020;97(4):631‐635. [DOI] [PubMed] [Google Scholar]
- 5. Lim J, Hwang IC, Choi HM, Yoon YE, Cho GY. Comparison of cardiovascular and renal outcomes between dapagliflozin and empagliflozin in patients with type 2 diabetes without prior cardiovascular or renal disease. PLoS One. 2022;17(10):e0269414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Szalat A, Perlman A, Muszkat M, Khamaisi M, Abassi Z, Heyman SN. Can SGLT2 inhibitors cause acute renal failure? Plausible role for altered glomerular hemodynamics and medullary hypoxia. Drug Saf. 2018;41:239‐252. [DOI] [PubMed] [Google Scholar]
- 7. Vasilakou D, Karagiannis T, Athanasiadou E, et al. Sodium‐ glucose cotransporter 2 inhibitors for type 2 diabetes: a systematic review and meta‐analysis. Ann Intern Med. 2013;159:262‐274. [DOI] [PubMed] [Google Scholar]
- 8. U.S. Food and Drug Administration . FDA drug safety communication: FDA strengthens kidney warnings for diabetes medicines canagliflozin (Invokana, Invokamet) and dapagliflozin (Farxiga, Xigduo XR). 2016. Accessed April 21, 2024. https://www.fda.gov/Drugs/DrugSafety/ucm505860.htm
- 9. Perlman A, Heyman SN, Matok I, Stokar J, Muszkat M, Szalat A. Acute renal failure with sodium‐glucose‐cotransporter‐2 inhibitors: analysis of the FDA adverse event report system database. Nutr Metab Cardiovasc Dis. 2017;27(12):1108‐1113. [DOI] [PubMed] [Google Scholar]
- 10. Johannes CB, Beachler DC, Layton JB, et al. Post‐authorization safety study of hospitalization for acute kidney injury in patients with type 2 diabetes exposed to dapagliflozin in a real‐world setting. Drug Saf. 2023;46(2):157‐174. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Mosenzon O, Wiviott SD, Cahn A, et al. Effects of dapagliflozin on development and progression of kidney disease in patients with type 2 diabetes: an analysis from the DECLARE‐TIMI 58 randomised trial. Lancet Diabetes Endocrinol. 2019;7(8):606‐617. [DOI] [PubMed] [Google Scholar]
- 12. Han E, Kim A, Lee SJ, et al. Characteristics of dapagliflozin responders: a longitudinal, prospective, nationwide dapagliflozin surveillance study in Korea. Diabetes Ther. 2018;9(4):1689‐1701. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Park S, Kim HJ, Won H, Lee HE, Cho H, Choi NK. Comparison of safety profiles for dapagliflozin based on EMA and FDA safety issues: challenges and future of post‐marketing surveillance in Korea. PLoS One. 2024;19(11):e0314363. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Horii T, Oikawa Y, Kunisada N, Shimada A, Atsuda K. Acute kidney injury in Japanese type 2 diabetes patients receiving sodium–glucose cotransporter 2 inhibitors: a nationwide cohort study. J Diabetes Investig. 2022;13(1):42‐46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Chung MC, Hung PH, Hsiao PJ, et al. Sodium‐glucose transport protein 2 inhibitor use for type 2 diabetes and the incidence of acute kidney injury in Taiwan. JAMA Netw Open. 2023;6(2):e230453. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Kyoung DS, Kim HS. Understanding and utilizing claim data from the Korean National Health Insurance Service (NHIS) and Health Insurance Review & Assessment (HIRA) database for research. J Lipid Atheroscler. 2022;11(2):103‐110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Zhao M, Sun S, Huang Z, Wang T, Tang H. Network meta‐analysis of novel glucose‐lowering drugs on risk of acute kidney injury. Clin J Am Soc Nephrol. 2020;16(1):70‐78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Mitsuboshi S, Morizumi M, Kotake K, Kaseda R, Narita I. Individual dipeptidyl peptidase‐4 inhibitors and acute kidney injury in patients with type 2 diabetes: a systematic review and network meta‐analysis. Basic Clin Pharmacol Toxicol. 2024;135(1):71‐80. [DOI] [PubMed] [Google Scholar]
- 19. Choi NK, Chang Y, Jung SY, et al. A population‐based case‐crossover study of polyethylene glycol use and acute renal failure risk in the elderly. World J Gastroenterol: WJG. 2011;17(5):651‐656. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Cho H, Lim E, Kim HJ, Jeong NY, Choi NK. Association between influenza vaccination and acute kidney injury among the elderly: a self‐controlled case series. Pharmacoepidemiol Drug Saf. 2024;33(9):e70006. [DOI] [PubMed] [Google Scholar]
- 21. Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD‐9‐CM and ICD‐10 administrative data. Med Care. 2005;43(11):1130‐1139. [DOI] [PubMed] [Google Scholar]
- 22. Kaur A, Sharma GS, Kumbala DR. Acute kidney injury in diabetic patients: a narrative review. Medicine. 2023;102(21):e33888. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Schetz M, Dasta J, Goldstein S, Golper T. Drug‐induced acute kidney injury. Curr Opin Crit Care. 2005;11(6):555‐565. [DOI] [PubMed] [Google Scholar]
- 24. Pannu N, Nadim MK. An overview of drug‐induced acute kidney injury. Crit Care Med. 2008;36(4):S216‐S223. [DOI] [PubMed] [Google Scholar]
- 25. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet. 2007;370(9596):1453‐1457. [DOI] [PubMed] [Google Scholar]
- 26. Nadkarni GN, Ferrandino R, Chang A, et al. Acute kidney injury in patients on SGLT2 inhibitors: a propensity‐matched analysis. Diabetes Care. 2017;40(11):1479‐1485. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Rampersad C, Kraut E, Whitlock RH, et al. Acute kidney injury events in patients with type 2 diabetes using SGLT2 inhibitors versus other glucose‐lowering drugs: a retrospective cohort study. Am J Kidney Dis. 2020;76(4):471‐479. [DOI] [PubMed] [Google Scholar]
- 28. Güzel C, Yeşiltaş S, Daşkaya H, Uysal H, Sümer I, Türkay M. The effect of gender on acute kidney injury developing in the intensive care unit. Hippokratia. 2019;23(3):126. [PMC free article] [PubMed] [Google Scholar]
- 29. Wang Y, Huang X, Xia S, et al. Gender differences and risk factors for acute kidney injury following cardiac surgery: a single center retrospective cohort study. Heliyon. 2023;9(12):e22177. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Loutradis C, Pickup L, Law JP, et al. Acute kidney injury is more common in men than women after accounting for socioeconomic status, ethnicity, alcohol intake and smoking history. Biol Sex Differ. 2021;12:1‐12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Heerspink HJ, Stefánsson BV, Correa‐Rotter R, et al. Dapagliflozin in patients with chronic kidney disease. N Engl J Med. 2020;383(15):1436‐1446. [DOI] [PubMed] [Google Scholar]
- 32. Heerspink HJ, Cherney D, Postmus D, et al. A pre‐specified analysis of the dapagliflozin and prevention of adverse outcomes in chronic kidney disease (DAPA‐CKD) randomized controlled trial on the incidence of abrupt declines in kidney function. Kidney Int. 2022;101(1):174‐184. [DOI] [PubMed] [Google Scholar]
- 33. Mende CW. Chronic kidney disease and SGLT2 inhibitors: a review of the evolving treatment landscape. Adv Ther. 2022;39(1):148‐164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Mavrakanas TA, Tsoukas MA, Brophy JM, Sharma A, Gariani K. SGLT‐2 inhibitors improve cardiovascular and renal outcomes in patients with CKD: a systematic review and meta‐analysis. Sci Rep. 2023;13(1):15922. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Yu MK, Vart P, Jongs N, et al. Effects of dapagliflozin in chronic kidney disease across the spectrum of age and by sex. J Gen Intern Med. 2024;39(6):921‐930. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Cahn A, Melzer‐Cohen C, Pollack R, Chodick G, Shalev V. Acute renal outcomes with sodium‐glucose co‐transporter‐2 inhibitors: real‐world data analysis. Diabetes Obes Metab. 2019;21(2):340‐348. [DOI] [PubMed] [Google Scholar]
- 37. Patorno E, Gopalakrishnan C, Franklin JM, et al. Claims‐based studies of oral glucose‐lowering medications can achieve balance in critical clinical variables only observed in electronic health records. Diabetes Obes Metab. 2018;20(4):974‐984. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1.–S2.
Data Availability Statement
The database is not available due to the privacy policy.