Influence of sodium/glucose cotransporter-2 inhibitors on the incidence of acute kidney injury: a meta-analysis

Background Sodium/glucose cotransporter-2 inhibitors (SGLT2i) are associated with cardiovascular benefits. The aim of this systematic review and meta-analysis is to summarize the influence of SGLT2i on the incidence of acute kidney injury (AKI), and to ascertain whether it is affected by confounding variables such as age, baseline renal function and concurrent use of renin-angiotensin-aldosterone system inhibitors (RAASi) or mineralocorticoid receptor antagonists (MRA). Methods PubMed, Embase, and Cochrane Library databases were searched for randomized controlled trials comparing the influence of SGLT2i versus placebo/blank treatment on AKI in the adult population. A fixed-effect model was used if the heterogeneity was not significant; otherwise, a randomized-effect model was used. Results Eighteen studies comprising 98,989 patients were included. Compared with placebo/blank treatment, treatment with SGLT2i significantly reduced the risk of AKI (risk ratio [RR]: 0.78, 95% confidence interval [CI]: 0.71 to 0.84, p < 0.001; I 2 = 0%). Subgroup analysis suggested consistent results in patients with diabetes, chronic kidney disease, and heart failure (for subgroup difference, p = 0.32). Finally, univariate meta-regression suggested that the influence of SGLT2i on the risk of AKI was not significantly modified by variables such as age (coefficient: 0.011, p = 0.39), baseline estimated glomerular filtration rate (coefficient: −0.0042, p = 0.13) or concomitant use of RAASi (coefficient: 0.0041, p = 0.49) or MRA (coefficient: −0.0020, p = 0.34). Conclusion SGLT2i may be effective in reducing the risk of AKI, and the effect might not be modified by age, baseline renal function and concurrent use of RAASi or MRA.


Introduction
Sodium/glucose cotransporter-2 inhibitors (SGLT2i) represent a novel class of oral antidiabetic medications that have demonstrated additional advantageous effects on cardiac and renal function (Frak et al., 2023;Klen and Dolzan, 2023;Lam-Chung, 2023).From a pharmacological standpoint, SGLT2i functions by inhibiting the reabsorption of glucose in the initial proximal tubule of the kidney, thereby augmenting the excretion of glucose in the urine and reducing the overall glucose burden on the body (Vallon and Verma, 2021).In individuals diagnosed with type 2 diabetes mellitus (T2DM), an initial meta-analysis of three extensive clinical trials revealed that the utilization of SGLT2i was associated with an 11% decrease in the risk of major adverse cardiovascular events, a 23% decrease in the risk of cardiovascular death or hospitalization for heart failure (HF), and a 45% decrease in the risk of progression of renal disease (Zelniker et al., 2019).In a study involving patients with HF, it was demonstrated that SGLT2i effectively reduced the likelihood of cardiovascular death and hospitalizations for HF across a diverse range of patients, thus establishing their significance as a fundamental therapy for HF, regardless of ejection fraction or care setting (Vaduganathan et al., 2022).Furthermore, a recent meta-analysis encompassing 13 clinical trials revealed that SGLT2i exhibited efficacy in altering the risk of kidney disease progression, not only in patients with T2DM at high cardiovascular risk, but also in patients with chronic kidney disease (CKD) or HF regardless of diabetic status (2022).Consequently, the indications for SGLT2i have expanded beyond T2DM to include HF and CKD, supported by accumulating evidence.Nevertheless, conflicting findings have emerged regarding the potential occurrence of acute kidney injury (AKI) when utilizing SGLT2i (Copur et al., 2023).One case report documented a dialysis-dependent AKI following the initiation of SGLT2i, with a suggested association to osmotic nephropathy (Phadke et al., 2020).Furthermore, a recent investigation utilizing the most up-to-date records from the United States Food and Drug Administration's Adverse Event Reporting System has indicated a potential link between SGLT2i and the development of AKI, although this association may be mitigated in instances where renin-angiotensin-aldosterone system inhibitors (RAASi), such as angiotensin converting enzyme inhibitors (ACEI) or angiotensin II receptor blockers (ARB), are concurrently administered (Katsuhara and Ikeda, 2021).Nevertheless, it is worth noting that several observational studies have failed to demonstrate an elevated risk of AKI associated with the use of SGLT2 inhibitors (Rampersad et al., 2020;Zhuo et al., 2022).In order to conduct a comprehensive assessment of the impact of SGLT2i on the occurrence of AKI, we conducted a systematic review and meta-analysis of eligible randomized controlled trials (RCTs).Furthermore, we investigated whether the effect of SGLT2i on the risk of AKI could be influenced by study-specific factors such as age, baseline renal function, and concurrent use of RAASi at the study level.

Methods
This study was designed and implemented according to the Cochrane Handbook guidelines (Higgins et al., 2021) and the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement (Page et al., 2021a;Page et al., 2021b).

Study selection
Studies were included if they fulfilled the following criteria according to the PICOS principles.P (patients): adult patient population without limitations of the diagnosis, which could be patients with T2DM, CKD, or HF.I (intervention): SGLT2i.C (control): placebo or blank treatment.O (outcome): incidence of AKI compared between patients with SGLT2i and controls during follow-up.The diagnosis of AKI was in accordance with the criteria used among the original studies S (study design): parallel-group RCTs, published as full-length articles in peer-reviewed journals.
Non-RCTs, studies that did not include an intervention group of SGLT2i, those comparing the effects of different doses of SGLT2i, single-arm studies without controls, or studies not evaluating the outcome of AKI were excluded.For studies with overlapping patients, the study with the largest sample size was included.

Data extraction and quality assessment
The process of data extraction, mining, and quality evaluations was carried out by two authors working independently.In the event of any disagreement, the corresponding author was consulted to address and resolve such inconsistencies.Information regarding publication details (author, year of publication, and study country), study design (blind or open-label), patient characteristics (diagnosis, demographic information, baseline renal function as evaluated by estimated glomerular filtration rate (eGFR), proportions of patients with concurrent use of any RAASi, and proportions of patients with concurrent use of mineralocorticoid receptor antagonists [MRA]), details of interventions and controls, follow-up durations, and diagnostic criteria for AKI was extracted.The quality of RCTs was assessed utilizing the Cochrane Risk of Bias Tool (Higgins et al., 2021), adhering to the subsequent criteria: (1) random sequence generation; (2) allocation concealment; (3) participant and staff blinding; (4) outcome assessor blinding; (5) presentation of incomplete outcome data; (6) reporting of selective results; and (7) identification of other potential biases.

Statistical analysis
The numbers of patients with AKI events and total numbers of patients allocated to the SGLT2i and control groups were extracted from the original reports.The influence of SGLT2i on the incidence of AKI in adult patients compared to control was summarized as risk ratio (RR) and corresponding 95% confidence intervals (CIs).The Cochrane Q test was performed (Higgins and Thompson, 2002).Heterogeneity was also estimated by calculating I 2 and I 2 > 50% suggested significant heterogeneity (Higgins et al., 2003).In the pooled analyses, a random-effects model was employed when significant heterogeneity was identified; alternatively, a fixed-effects model was utilized (Higgins et al., 2021).A sensitivity analysis was performed by only including high-quality studies (all seven domains of Cochrane Risk of Bias Tool judged as low risk).Additionally, a predefined subgroup analysis was conducted based on the patients' diagnosis and the specific SGLT2i drugs administered.Furthermore, a univariate meta-regression analysis was conducted to investigate whether the study characteristics of continuous variables could significantly alter the impact of SGLT2i on AKI, such as mean age of the patients, proportion of men, mean eGFR at baseline, proportion of patients using any RAASi, proportion of patients using MRA, and mean follow-up duration of the study.Publication bias was assessed using funnel plots and Egger's regression asymmetry test (Egger et al., 1997).Statistical significance was defined as p < 0.05.The statistical analysis was performed using Stata software (version 12.0; Stata Corporation) and RevMan (version 5.1; Cochrane, Oxford, United Kingdom).

Search results
A diagram illustrating the process of database searching and study identification is presented in Figure 1.The search of the databases yielded a total of 798 articles, of which 599 were identified as unique after removing duplicates.Subsequently, 557 articles were excluded based on their title and abstract, primarily due to their lack of relevance to the research objectives.A thorough examination of the full text was conducted on 42 articles, resulting in the exclusion of 24 articles for the reasons depicted in Figure 1.Ultimately, the final analysis encompassed a total of 18 RCTs (Zinman et al., 2015;Neal et al., 2017;McMurray et al., 2019;Perkovic et al., 2019;Wiviott et al., 2019;Cannon et al., 2020;Heerspink et al., 2020;Bhatt et al., 2021a;Anker et al., 2021;Bhatt et al., 2021b;Kosiborod et al., 2021;Zannad et al., 2021;Solomon et al., 2022;Voors et al., 2022;Feitosa et al., 2023;Herrington et al., 2023;Butler et al., 2024;Cox et al., 2024).

Publication bias
The symmetrical funnel plots observed in the meta-analyses of the impact of SGLT2i on AKI in adult patients indicate a minimal likelihood of publication bias (Figure 5).Furthermore, the results of Egger's regression test support this notion, as it yielded a p-value of 0.32, indicating a low risk of publication bias.

Discussion
In this meta-analysis, we conducted a comprehensive synthesis of data from 18 RCTs, comprising 20 datasets.The findings of our study indicate a significant reduction in the risk of AKI among adult patients when comparing the use of SGLT2i to placebo or blank treatment.The sensitivity analysis limited to high-quality RCTs showed similar results.Subgroup analyses further demonstrated consistent results in patients with T2DM, CKD, and HF.Additionally, our subgroup analysis suggests that the impact of SGLT2i on AKI does not appear to be influenced by the specific type of SGLT2i utilized.Finally, meta-regression analysis suggested that the influence of SGLT2i on AKI was also not likely to be modified by difference of study characteristics, such as mean age of the patients, proportion of men, mean baseline eGFR, proportions of patients with concomitant use of RAASi and MRA, or follow-up durations.
Taken together, these results indicate that SGLT2i may be effective in reducing the risk of AKI, and the effect of SGLT2i on AKI may not be influenced by the baseline renal function or concurrent use of RAASi.
Some meta-analyses were published before or during the preparation this manuscript, which generally showed that SGLT2 inhibitors can exert the benefit in reducing AKI in patients with T2D, heart failure, or CKD; and this benefit does not vary with various characteristics, such as the diagnosis of the patients and type of SGLT2 inhibitors (Menne et al., 2019;Neuen et al., 2019;Zhao et al., 2020;Qiu et al., 2021;Baigent et al., 2022;Gong et al., 2022;Rigato et al., 2023).Compared to the previous meta-analyses, our study has several strengths.First, an extensive literature search was performed which retrieved 18 relevant up-to-date RCTs.Second, only RCTs were included, which minimized the biases related to the design of observational studies.In addition, although the results of the overall and subgroup analyses were generally consistent with the findings of the previous metaanalyses, we for the first time performed meta-regression analyses to investigate the potential influence of study characteristics such as age, baseline renal function, and concurrent use of RAASi at the study level.This is clinically important, because these factors have been related to the risk of AKI.Overall, results of the meta-analysis provided further evidence that SGLT2i may be effective in reducing the risk of AKI, and the effect might not be modified by age, baseline renal function and concurrent use of RAASi or MRA.
Although concerns have been raised reading AKI related to SGLT2i use in some case reports, subsequent investigations in high Forest plots for the subgroup analysis of the influence of SGLT2i on the risk of AKI according to individual SGLT2i drugs used.
quality clinical trials and meta-analysis showed that SGLT2i may confer renal proactive efficacy and delay the deterioration of renal function (Lin et al., 2023).The current meta-analysis, by integrating the evidence from RCTs, further expanded the renal benefits of SGLT2i by showing that SGLT2i are effective in reducing the risk of AKI as compared to placebo/blank treatment.The potential mechanisms underlying the renal protective effect of SGLT2i may be multifactorial.An initial investigation conducted on non-diabetic mice using a renal ischemia/reperfusion injury model demonstrated that Luseogliflozin effectively mitigated peritubular capillary congestion/hemorrhage, alleviated hypoxia, and enhanced the expression of vascular endothelial growth factor (VEGF)-A, thereby exhibiting a protective effect on the kidneys during acute situations (Zhang et al., 2018).Furthermore, another study conducted on diabetic rats with myocardial infarction-associated AKI revealed that pretreatment with empagliflozin for 2 weeks resulted in improved hyperglycemia, elevated blood βhydroxybutyrate levels, suppressed expression of NGAL and KIM-1 induced by MI, and ultimately prevented the pathogenesis of AKI (Kuno et al., 2020).Furthermore, previous research has demonstrated the significant reduction of both systemic and renal inflammation by empagliflozin, which has contributed to the observed survival benefits in an LPS-model of acute septic renal injury (Maayah et al., 2021).Additionally, a more recent study has indicated that dapagliflozin may mitigate contrast-induced acute kidney injury through the suppression of the hypoxia-inducible factor-1α pathway (Huang et al., 2022).Consequently, there is a need for further investigation into the key molecular pathways that underlie the preventive effectiveness of SGLT2i on AKI.
Results of subgroup analysis suggested that although no significant difference was observed for the influence of each individual SGLT2i drugs on AKI, the positive results were mainly driven by studies involving empagliflozin, canagliflozin, and dapagliflozin, but not for studies with ertugliflozin or sotagliflozin.However, these results should be interpreted with caution because only two datasets were available for the subgroups of ertugliflozin and sotagliflozin, and more studies are needed for further evaluation.Interestingly, results of metaregression analysis suggested that the effect of SGLT2i on AKI did not seem to be significantly affected by eGFR at baseline, suggesting that potential renal protective efficacy of SGLT2i may also be consistent in patients with renal dysfunction before treatment (eGFR as low as 20 mL/min/1.73m 2 ).In addition, it has been suggested that excessive decline by SGLT2i combined with the excessive decline in trans-glomerular pressure induced by concomitant use of RAASi may further increase the risk of AKI (Szalat et al., 2018).Accordingly, we explored the influence of proportions of patients with concurrent use of RAASi and MRA on the effect of SGLT2i on AKI.Results suggested that the potential renal protective efficacy of SGLT2i may not be significantly modified by concurrent use of RAASi or MRA.This is consistent with a recently published post hoc analysis which showed that dapagliflozin consistently reduced the risk of kidney outcomes in T2DM patients irrespective of background use of various cardiovascular medications (Oyama et al., 2022).However, our results of metaregression analysis according to baseline renal function and concurrent use of RAASi should be considered as exploring study because these results were based on the analysis of study-level data rather than individual-patient data.
This meta-analysis also has limitations.First, different SGLT2i drugs with different dosages were used among the included studies.Further studies are needed to determine if the influence of SGLT2i on AKI is consistent among individual SGLT2i drugs, and if there is a  Funnel plots for the publication bias underlying the meta-analysis of the influence of SGLT2i on the risk of AKI in adult patients.
dose-effect relationship.Second, key aspects such as diabetes severity and duration, CKD, or HF, which have potential implications on SGLT2i effectiveness, may affect the influence of SGLT2i on AKI.Although, our meta-analysis is based on data at the study level rather than individual patient level; therefore, we were unable to determine the influence of these factors on the results.In addition, there are other medications which may also affect the risk of AKI besides RAASi and MRA, such as nonsteroidal anti-inflammatory drugs (NSAIDs).However, the status of NSAIDs use was generally not reported among the included studies, and we were therefore unable to determine its influence on the results of the meta-analysis.Moreover, we only included studies published in English as fulllength article in peer-reviewed journals.Grey literature, such as conference abstracts and unpublished data were not included.
Although excluding grey literature may improve the reliability of the finding because most grey literature are not strictly peer-reviewed, excluding these data may also increase the risk of publication bias.
Finally, for most of the included studies, AKI was diagnosed based on MDRA preferred term for AKI.The influence of different diagnostic criteria for AKI, particularly those applicable in real-world clinical practice needs to be further evaluated.As a summary, results of the meta-analysis suggest that SGLT2i may be effective in reducing the risk of AKI as compared to placebo/ blank treatment in adult patients, and the influence of SGLT2i on AKI may not be affected by baseline renal function and concurrent use of RAASi.

FIGURE 1
FIGURE 1Flowchart of literature search.

TABLE 1
Characteristics of the included studies.

TABLE 1 (
Continued) Characteristics of the included studies.

TABLE 2
Study quality evaluation via the Cochrane Risk of Bias Tool.