2016 FDA warning conflicts with recent evidence showing no increased risk for kidney injury with SGLT2 inhibitors following use of the treatment.
To date, countless studies have been published detailing the many benefits of sodium-glucose cotransporter-2 (SGLT-2) inhibitors use in diabetes. From the fewer adverse cardiovascular outcomes to the lessened mortality, SLGT-2 inhibitors are proving to be a good option for the treatment of diabetes.
However, where there are positives, negatives are sure to be found. The US Food and Drug Administration has voiced some concern with the use of this drug class, specifically the risk of acute kidney injury (AKI) in patients treated with either canagliflozin or dapagliflozin. In June 2016, the FDA strengthened an existing warning for these two drugs. Additions included warning label revision and guidance to minimize AKI risk. These warnings came as a result of 101 confirmed cases of AKI due to the use of these agents.
While the AKI warning is critical, healthcare professionals must always weigh the risks versus benefits of these drugs and consider patient-specific factors before prescribing these agents. Despite these warnings, recent studies EMPA-REG OUTCOME (empagliflozin) and CANVAS (canagliflozin) showed that SGLT-2 inhibitors can decrease risk of chronic kidney disease. Given this discrepancy, the investigators of this study wanted to consider the “real” world risk of AKI when SGLT-2 inhibitors (i.e. canagliflozin, dapagliflozin, and empagliflozin) are used in people with type 2 diabetes at two large healthcare systems.
Two study cohorts were used, Mount Sinai Health System (MSHS) and Geisinger Health System (GHS). Patients, identified through electronic medical records, were included if they had a type 2 diabetes diagnosis and available serum creatinine data. Each cohort was separated into either the control group (i.e. nonuser of SGLT-2 inhibitors) or treatment group (i.e. user of SGLT-2 inhibitor). Additional cohort details can be found in Table 1.
|Table 1: Description of Propensity-Matched Cohorts|
|Cohort||Period of Duration||Control Arm (N)||Treatment Arm (N)|
|Mount Sinai Health System||January 1, 2014 – December 30, 2016||372||372|
|Geisinger Health System||January 1, 2013 – February 10, 2017||1,207||1,207|
The primary outcome of this study was to identify AKI events after index date (i.e. for users = first date an SGLT2 inhibitor was prescribed; nonuser = creatinine measurement date after 2013) in the inpatient setting. Key definitions, including AKI event classification, baseline creatinine, and GSH specific AKI definitions, can be found in Table 2.
|Table 2: AKI Definitions|
|GSH-Specific AKI Event Classification||AKI
|1. KDIGO creatinine-based criteria, or
2. Serum creatinine criteria (i.e. increase serum creatinine by ≥ 0.3 mg/dL within 48h or increase in serum creatinine by ≥ 1.5 times baseline value in the prior 7 days).
|1. Outpatient creatinine prior to AKI episode, or
2. Average creatinine over the past year prior to AKI episode. creatinine value within 10 days of AKI event).
|1. KDIGO creatinine-based criteria, or
2. Increase in serum creatinine by 50% from outpatient baseline measured one year prior to admission, or
3. Increase in serum creatinine by 50% over 7 days in inpatient setting
|1. Change in serum creatinine from baseline during AKI event,
2. Peak serum creatinine during AKI event.
After propensity matching, the MSHS cohort users were predominantly on canagliflozin (71.8%) followed by dapagliflozin (19.4%) and empagliflozin (8.9%). Users and nonusers were matched for race, hemoglobin, HbA1c, thiazide diuretic use, and metformin use. In the GHS cohort, like the MSHS cohort, users were also mainly on canagliflozin. After propensity matching, users and nonusers were matched, however they were not matched in the HbA1c, metformin use, antihypertensive use, loop diuretic use, and thiazide diuretic use.
Following propensity matching, in the MSHS cohort, 3.8% of users experienced an AKI event compared to 9.7% of nonusers (p = 0.002). Severity of AKI events were noted by changes from baseline and peak serum creatinine values. Changes in baseline for users and nonusers was 0.5 mg/dL (IQR 0.4-0.7) and 0.9 mg/dL (IQR 0.8-1.3; p = 0.004). Peak serum creatinine values for users and nonusers was 1.6 mg/dL (IQR 1.4-1.8) and 1.9 mg/dL (IQR 1.6-2.4; p = 0.02). AKI events requiring acute dialysis occurred in one patient for both groups and 7 users (out of 14) had their SGLT-2 inhibitor discontinued without restart. Separately, in the GSH cohort, 2.2% of users experienced AKI events compared to 4.6% of nonusers (p < 0.01). No difference in baseline serum creatinine changes (0.6 vs 0.6 mg/dL; p = 0.8) or peak serum creatinine measurements (1.7 vs 1.6 mg/dL; p = 0.9), was found for user and nonusers, respectively. AKI events requiring acute dialysis occurred in one nonuser, 3 users (out of 26) had their medication restarted after an event, and 23 users had their medication discontinued with restart.
Sensitivity analyses using AKIICD-specific events showed consistent results in the MSHS cohort and showed comparable results in the GHS cohort (1.2% user AKI vs 3.0% nonuser AKI rate; p < 0.01).
The last endpoint to be analyzed was the association between SGLT-2 inhibitor use and AKI incidence. In the MSHS cohort, after adjustments were made (i.e. metformin use, HbA1c, smoking, thiazide use, race), an adjusted hazard ratio of 0.4 (95% CI 0.2-0.7; p = 0.004) was shown. In the GHS cohort, after adjustments were made (i.e. diastolic BP, total cholesterol, HbA1c, hemoglobin, albuminuria, antihypertensive use, loop diuretic use, thiazide diuretic use, and metformin use), an adjusted hazard ratio of 0.6 (95% CI 0.4-1.1; p = 0.09) was shown. Overall, the adjusted hazard ratios were both lower in the users in both cohorts, but the hazard ratio for the GHS cohort was not significant. Additionally, using the AKIICD-specific events revealed similar results, lower hazard ratios for AKI in the user group.
To summarize, the evidence presented in this propensity-matched cohort-driven study did not show an increased risk of AKI with SGLT-2 inhibitor use. In either cohort, data revealed fewer AKI events in patients prescribed SGLT-2 inhibitors. Instances of AKI severity were few and not as severe in those taking SGLT-2 inhibitors. Finally, those taking the SGLT-2 inhibitors known for having the highest risk for AKI, canagliflozin and dapagliflozin, show no increased risk of AKI. The results of this study show that SGLT-2 inhibitor use, when compared to nonusers, over one year within two health system cohorts has no increased risk of AKI.
- Currently the FDA has stated that two SGLT-2 inhibitors, canagliflozin and dapagliflozin, are known to increase the risk of acute kidney injury as evidenced by 101 confirmed cases.
- Nadkarni, et al study concluded that SGLT-2 inhibitor use did not show an increased risk of AKI.
- AKI severity, experienced by users of SGLT-2 inhibitors, was no worse than patients not taking SGLT-2 inhibitors.
Nadkarni, G., Ferrandino, R., Chang, A., Surapaneni, A., Chauhan, K., Poojary, P., Saha, A., Ferket, B., Grams, M., and Coca, S. Acute Kidney Injury in Patients on SGLT2 Inhibitors: A Propensity-Matched Analysis. Diabetes Care. 40.11 (2017): 1479-1485. https://doi.org/10.2337/dc17-1011.
FDA Drug Safety Communication: FDA strengthens kidney warnings for diabetes medicines canagliflozin (Invokana, Invokamet) and dapagliflozin (Farxiga, Xigduo XR). U.S. Food & Drug Administration. 14 June 2016. Available from: https://www.fda.gov/Drugs/DrugSafety/ucm505860.htm
Kaytie A. Weierstahl, Pharm.D. Candidate, LECOM School of Pharmacy