Type 2 diabetes mellitus (T2DM) is one of the most common factors leading to renal failure and lacks effective prevention and treatment methods (Saran et al., 2017; Warren et al., 2018). The number of patients receiving renal replacement therapy for renal failure is reported to exceed 5 million worldwide in 2030 (Liyanage et al., 2015). Therefore, effective measures to improve the renal prognosis of patients with T2DM have become an area of great concern. In recent years, sodium–glucose cotransporter-2 (SGLT-2) inhibitors, a new class of oral hypoglycemic drugs that include canagliflozin, empagliflozin, and dapagliflozin, have shown renal benefits independent of hypoglycemic effects both in outcomes research and in clinical practice (Vlado et al., 2018; Heerspink et al., 2020a; De Nicola et al., 2020; Jongs et al., 2021). These improvements mainly include reducing proteinuria and delaying the deterioration of renal function (Ma et al., 2023). The CREDENCE trial, the first clinical trial to report the impact of SGLT-2 inhibitors on the kidneys, showed that the risk of renal failure was lower in the canagliflozin group than in the placebo group among patients with T2DM combining chronic kidney disease (CKD) (Perkovic et al., 2019). The subsequent EMPEROR trial showed that empagliflozin slowed the rate of renal function decline both in patients with and without CKD regardless of the severity of renal injury at baseline (Zannad et al., 2021). A large phase III clinical study, DAPA-CKD, showed that dapagliflozin treatment could delay the progression of kidney disease and reduce the risk of the main endpoint (composite of a sustained decline in the estimated glomerular filtration rate (eGFR) of at least 50%, end-stage kidney disease, or death from renal or cardiovascular causes) by 5.3%, including in patients without diabetes (Heerspink et al., 2020a). Thus, based on such a large number of clinical trial results, multiple countries have successively approved SGLT-2 inhibitors for the treatment of CKD (Herrington et al., 2021; Administration USFD Home-News, 2024; Agency, 2024; Astrazeneca, 2024).

Multiple mechanisms have been reported to be involved in the protective effect of SGLT-2 inhibitors on the kidneys (Heerspink et al., 2016). eGFR usually increases with the development of diabetes, which is an indicator of the development of diabetic nephropathy. SGLT-2 inhibitors reduce glucose levels by inhibiting the re-absorption of glucose and sodium in the proximal tubules, increasing their transport to the distal renal tubules, normalizing tubular bead feedback signals, and alleviating over filtration (Vallon and Thomson, 2020). The increase in glomerular perfusion leads to an increase in vasoconstriction of afferent arterioles, followed by a decrease in glomerular perfusion and intraglomerular pressure. These changes rapidly reduce the eGFR, which, however, gradually stabilizes over time (Wanner et al., 2016). This phenomenon was confirmed by the CREDENCE and DAPA-CKD trials (Perkovic et al., 2019; Heerspink et al., 2020a). Moreover, Zannad et al. found that the eGFR may decrease slightly following SGLT-2 inhibitors administration, which does not affect renal function (Zannad et al., 2022).

However, the research included in these clinical trials is highly population selective, which cannot directly reflect the real-world trend of the renal function change over time in patients with T2DM after SGLT-2 inhibitors administration. Limited studies have reported the impact of SGLT-2 inhibitors on eGFR in the real world population, (Fadini et al., 2018; Kosiborod et al., 2018; Udell et al., 2018; Fadini et al., 2019; Kinguchi et al., 2019; Lin et al., 2019; Sezai et al., 2019; Heerspink et al., 2020b; Pasternak et al., 2020), with only a few having reported specific renal test results, such as eGFR and urinary albumin/creatinine ratio (UACR) (Heerspink et al., 2020b; Lin et al., 2021). Therefore, this study aims to conduct a retrospective analysis of the real-world data of outpatients with T2DM using SGLT-2 inhibitors with different baseline renal functions in mainland China. It is expected that this study will extend the existing evidence from the clinical trials to a wider and more heterogeneous population to provide further support, and a well-informed rationale behind the clinical application of SGLT-2 inhibitors, while ensuring that patients use these more safely and effectively.

This was the first real-world study to focus on renal protection of SGLT-2 inhibitors in patients with T2DM with different baseline renal functions, conducted in mainland China. Although the data was only from the outpatient prescription database of one hospital, Peking University People’s Hospital is one of the first hospitals in mainland China to use SGLT-2 inhibitors drugs. Moreover, adequate outpatient flow ensures population diversity, and that the results are applicable to a wider Asian population. This study further validated the protective effect of SGLT-2 inhibitors on renal function in T2DM patients based on real-world records.

Until now, only a few studies have reported the effect of SGLT-2 inhibitors on renal prognosis in patients with different baseline renal functions. The proportions of patients in different subgroups in our study were similar to that in the CVD-REAL trial, which was the first largest transnational observational cohort study to evaluate the effects of SGLT-2 inhibitors treatment on kidney prognosis of patients in the real world (Heerspink et al., 2020b). However, the baseline eGFR of patients in the CVD-REAL trial was significantly higher than that of our study. A cohort study from Taiwan, China, explored the effects of SGLT-2 inhibitors on the renal prognosis of Asian patients with T2DM in the real world (Lin et al., 2021). The average baseline eGFR level of patients included in Taiwanese study was similar to that of our study, however the population composition based on baseline eGFR grouping was significantly different. The proportion of patients with atherosclerosis cardiovascular disease reached 51.7% in our study, significantly higher than that in the former studies (Heerspink et al., 2020b; Lin et al., 2021). In addition, the 18th month follow-up rates of the three subgroups in our study were higher than those of other real-world studies (Lin et al., 2019; Heerspink et al., 2020b; Lin et al., 2021). Although there were differences in patient characteristics and study design between our study and previous studies (Heerspink et al., 2016; Heerspink et al., 2020b; Lin et al., 2021), the overall trend of eGFR changes in patients using SGLT-2 inhibitors was consistent; the decrease in eGFR reached a maximum value in the first month and then began to rise gradually and tended to return to baseline level. However, some studies have shown that SGLT-2 inhibitors have the advantage of slowing the downward trend of eGFR, but do not show an upward trend (Wheeler et al., 2021; Zannad et al., 2021; Wheeler et al., 2022). In summary, the results of our research provided the support for the improvement of renal function by SGLT-2 inhibitors in the real world.

The trend of eGFR variation in the subgroup with eGFR ≤60 mL/min/1.73 m² in most studies was consistent with the overall results of the study (Chertow et al., 2021; Lin et al., 2021). However, in our study, the trend of eGFR changes during follow-up in this subgroup was significantly different from the overall results, while the results of other subgroups were similar to those of previous studies (Chertow et al., 2021; Lin et al., 2021). The average baseline eGFR in the eGFR ≤60 mL/min/1.73 m² subgroup in our study was 47.57 mL/min/1.73 m², and the average eGFR within 9 months after taking SGLT-2 inhibitors was higher than the baseline value, and then began to decline. Another study showed similar results, in eGFR ≤60 mL/min/1.73 m² subgroup, which showed that the eGFR was greater than the baseline value within 2 years after treatment with dapagliptin, and then began to decline (Mosenzon et al., 2019). In addition, the CVD-REAL study reported that in the corresponding subgroup (n = 973), the eGFR of patients after taking SGLT-2 inhibitors increased by 0.22 mL/min/1.73 m², however, the average value of eGFR in the subgroup and its change over time were not reported (Heerspink et al., 2020b). It is possible that differences in renal function among patients included in the studies led to different conclusions. Moreover, considering that not all patients had a corresponding eGFR value at each time point during the follow-up period and the number of patients included in the three subgroups was significantly different, careful interpretation of the results is required. Thus, increasing the sample size of the eGFR ≤60 mL/min/1.73 m² subgroup may increase robustness in this study. In summary, our results indicated that SGLT-2 inhibitors have a certain improvement effect on renal function in T2DM patients with eGFR ≤60 mL/min/1.73 m² over a certain period of time.

The differences in SGLT-2 inhibitors action on renal function may also be related to follow-up time. Some studies have shown that eGFR decreases sharply in a dose-dependent manner within a few weeks, (Heerspink et al., 2016; Lin et al., 2019), while the probability of kidney-related events increases with time (Lin et al., 2019). However, the potential mechanism of SGLT-2 inhibitors in protecting the kidneys is not fully understood. The results of the EMPA-REG trial showed that the potential change in glomerular pressure contributes to the improved renal end point (Wanner et al., 2016). Several studies report that the renal hemodynamic mechanism mediated by natriuretic and glucose-induced osmotic diuresis is the main reason for the short-term biphasic changes in renal function caused by SGLT-2 inhibitors (Heerspink et al., 2018; Cherney et al., 2020; van Bommel et al., 2020). Moreover, the type of SGLT-2 inhibitors also affects the degree of eGFR decrease in patients. Huilin et al. pointed out that empagliflozin causes fewer kidney-related events than dapagliflozin, (Tang et al., 2017), however, Lin et al. reported contradictory results (Lin et al., 2019). Thus, in future, research should focus on the correlation between SGLT-2 inhibitors type and the trend of eGFR changes.

Our study also analyzed the UACR trend over time after using SGLT-2 inhibitors in patients with T2DM. Although the sample size for UACR was significantly smaller than that of eGFR, we still performed subgroup analysis according to the baseline value. During the follow-up period, patients with baseline proteinuria showed varying degrees of improvement. For patients with microalbuminuria at baseline, the maximum reduction of UACR after taking SGLT-2 inhibitors reached 38.9%. Patients with clinical proteinuria at baseline showed the maximum decrease in UACR during the observation period, which was 47.5%. For patients with proteinuria in our study, the degree of reduction in proteinuria caused by SGLT-2 inhibitors was similar to the results of large randomized controlled trials, ranging from 18% to 40%. (Vlado et al., 2018; Perkovic et al., 2019; Jongs et al., 2021). The prespecified analysis of the DAPA-CKD trial showed that SGLT-2 inhibitors can reduce proteinuria in patients with diabetes. (Pollock et al., 2019; Heerspink et al., 2020a; Jongs et al., 2021). Our study further confirmed the role of SGLT-2 inhibitors in reducing proteinuria in patients with T2DM. Moreover, Jongs et al. found a significant reduction of UACR during dapagliflozin treatment was related to eGFR decrease during the follow-up period. (Jongs et al., 2021). In future, the sample size needs to be increased when studying the correlation between eGFR and UACR.

Moreover, we also analyzed the impact of SGLT-2 inhibitors on HbA1c levels. Most of the patients included in our study were treated with SGLT-2 inhibitors in combination with other types of hypoglycemic drugs. We conducted statistical analyses on the proportion of patients whose HbA1c reached the standard value after initiation of SGLT-2 inhibitors and found that the overall compliance rate of HbA1c showed an upward trend after initiating SGLT-2 inhibitors treatment. After the follow-up period, the overall HbA1c decreased by an average of 1.1% compared to baseline. Similarly, other studies have shown that SGLT-2 inhibitors can reduce HbA1c in patients with T2DM by 0.11%–0.9% after a treatment cycle of not less than 6 months (Jabbour et al., 2014; Ridderstråle et al., 2014; Rodbard et al., 2019; Inzucchi et al., 2021).

In addition, our results further confirmed that SGLT-2 inhibitors can reduce UA levels in patients with high UA levels (Doehner et al., 2022; Ferreira et al., 2022; Suijk et al., 2022). During the follow-up period, the time point for reaching the maximum decrease, with the value of −57.97 μmol/L, was 1 month after initiating SGLT-2 inhibitors treatment in this subgroup, which was consistent with previous studies (Doehner et al., 2022). The mechanism of SGLT-2 inhibitors in reducing UA levels has been shown to be potentially different from that of UA-lowering drugs (McDowell et al., 2022). La et al. (2022) found that the reduction of UA by SGLT-2 inhibitors was related to its inhibition of low-grade inflammation. In addition, animal experiments showed that empagliflozin attenuates hyperuricemia by upregulation of ABCG2 via the AMPK/AKT/CREB signaling pathway in type 2 diabetic mice (Lu et al., 2020). Butt et al. (2023) suggested that the lowering effect of SGLT-2 inhibitors on UA can reduce the required dosage of urate-lowering therapy drugs. Although the mechanism of lowering UA by SGLT-2 inhibitors is not yet clear, this study further confirms the urate-lowering effect of SGLT-2 inhibitors in the real world.

We also investigated the effect of SGLT-2 inhibitors on blood lipids. We analyzed the blood lipid data of the remaining patients, particularly LDL-C, after excluding those with a history of using lipid modulating drugs. Currently, reports on the role of SGLT-2 inhibitors in LDL-C have been controversial (Fadini et al., 2018; Kinguchi et al., 2019; Sezai et al., 2019; Berkovic et al., 2020; Osto et al., 2023). In our study, although there was no significant statistical difference in LDL-C level between patients at the end of the follow-up period and the baseline, LDL-C level was significantly lower than the baseline value at 1, 3, and 6 months after taking SGLT-2 inhibitors, indicating that SGLT-2 inhibitors had a certain lipid modulating effect. Sezai et al. (2019) have suggested that the improvement of LDL-C levels may be related to the reduction of fat volume in patients taking SGLT-2 inhibitors. In addition, the SGLT-2 inhibitors effect on TC was similar to that of LDL-C, and there was a trend towards lower TG after SGLT-2 inhibitors usage. However, SGLT-2 inhibitors had almost no effect on HDL-C. Therefore, SGLT-2 inhibitors can reduce blood lipids, temporarily, with the effect gradually disappearing over time. Although the mechanism and effect of SGLT-2 inhibitors on blood lipids remains controversial, our retrospective study based on the real world excluded the interference of lipid modulating drugs, which has strong persuasiveness. Previous research has shown that SGLT-2 inhibitors can reduce oxidative stress associated with T2DM, improve insulin sensitivity and lead to better cardiovascular outcomes and renal advantages (Andreadi et al., 2023). Therefore, in T2DM patients, the use of SGLT-2 inhibitors improves cardiovascular outcomes and controls metabolic consequences. People has attached more attention to the importance of SGLT-2 inhibitors in the treatment of diabetes in patients with cardiovascular diseases. Overall, our study still provides a practical basis for the protective effect of SGLT-2 inhibitors on the cardiovascular system.

This study had some limitations. First, this study was conducted in a single hospital, thus the sample size included in the study was small, especially in the eGFR ≤60 mL/min/1.73 m² subgroup. Second, the relevant laboratory examinations depended on patient compliance, resulting in significant data loss during the follow-up period. In addition to the small number of patients with proteinuria or CKD, most doctors tended to examine eGFR but ignored UACR, which made it difficult to collect sufficient UACR data. Third, we did not conduct further research on the effect of SGLT-2 inhibitors type and dosage on renal function in patients. Also, a longer follow-up time is required to further study the long-term effects of SGLT-2 inhibitors on the kidneys. Fourth, we only focused on the clinical efficacy of SGLT-2 inhibitors and did not perform statistical analysis on any acute kidney injury or urinary tract infection that they may cause. Finally, no in-depth analysis was performed on some mixed factors that may affect the observation indicators.

Our results indicate that during the 18-month follow-up, SGLT-2 inhibitors can improve renal function in patients with T2DM regardless of their baseline renal function. Although eGFR may decline sharply in the initial stage of medication, it gradually improves with prolonged medication. Our study further demonstrates that SGLT-2 inhibitors can reduce proteinuria in patients with T2DM. On the basis of existing clinical trials and real-world studies, our findings provide further evidence for the use of SGLT-2 inhibitors and their renal benefits in Asian populations.