Effects of Sodium-Glucose Cotransporter 2 Inhibitors on Renal Outcomes in Patients with Type 2 Diabetes: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

This study was conducted to investigate the effects of sodium-glucose cotransporter 2 (SGLT2) inhibitors on individual renal outcomes in patients with type 2 diabetes. We searched MEDLINE, Embase, and the Cochrane Central Register of Controlled Trials from inception to September 2017 to identify randomized controlled trials comparing SGLT2 inhibitors with placebo or antidiabetic drugs and reporting any renal outcomes in patients with type 2 diabetes. Additionally, we identified 4 articles which were published after the predefined period to include relevant data. A meta-analysis was performed to calculate weighted mean differences (WMDs) and relative risks (RRs) with 95% confidence intervals (CIs) for each renal outcome. We included 48 studies involving 58,165 patients in the analysis. SGLT2 inhibitors significantly lowered urine albumin-to-creatinine ratio (UACR) (WMD, −14.64 mg/g; 95% CI, −25.15 to −4.12; P = 0.006) compared with controls. The UACR-lowering effects of SGLT2 inhibitors were greater with a higher baseline UACR. Overall changes in estimated glomerular filtration rate (eGFR) were comparable between two groups (WMD, 0.19 mL/min/1.73 m2; 95% CI, −0.44 to 0.82; P = 0.552). However, SGLT2 inhibitors significantly slowed eGFR decline in patients with a higher baseline eGFR and a longer duration of treatment. Compared with controls, SGLT2 inhibitors significantly reduced the risk of microalbuminuria (RR, 0.69; 95% CI, 0.49 to 0.97; P = 0.032), macroalbuminuria (RR, 0.49; 95% CI, 0.33 to 0.73; P < 0.001), and worsening nephropathy (RR, 0.73; 95% CI, 0.58 to 0.93; P = 0.012). In addition, the risk of end-stage renal disease was significantly lower in SGLT2 inhibitors than in controls (RR, 0.70; 95% CI, 0.57 to 0.87; P = 0.001). In conclusion, SGLT2 inhibitors had beneficial renal effects by lowering the risk of albuminuria development or progression and reducing the risk of end-stage renal disease compared with placebo or other antidiabetic drugs.

Data sources and search strategy. We searched MEDLINE, Embase, and the Cochrane Central Register of Controlled Trials to identify RCTs of SGLT2 inhibitors with full-text articles published from inception to September 2017, regardless of language and publication status. The search terms used for SGLT2 inhibitors were SGLT2 inhibitor or SGLT-2 inhibitor or canagliflozin or dapagliflozin or empagliflozin or ertugliflozin or ipragliflozin or luseogliflozin or remogliflozin or sergliflozin or tofogliflozin (Supplementary Appendix 3). In addition, we identified 4 articles 11,[20][21][22] which were published after the predefined period to include all relevant data. Study selection. The RCTs comparing SGLT2 inhibitors with placebo or other antidiabetic drugs with ≥ 12 weeks of study duration in type 2 diabetes were included. We eliminated duplicate publications of original RCT and screened titles and abstracts. Among them, we selected RCTs that reported at least one of the following renal outcomes: changes in urine albumin-to-creatinine ratio (UACR) or eGFR, and incident microalbuminuria, macroalbuminuria, doubling of serum creatinine, renal failure, ESRD, RRT, dialysis, or kidney transplantation. The selection criteria for renal outcomes have been reported elsewhere 23 . Pooled analyses or secondary analyses were included when they provided additional information about renal outcomes beyond that found in original RCT articles. We included two publications of canagliflozin trials 20,22 which were reported after the prespecified analysis considering the effect size and weight of these studies. Data extraction. Two authors (J.H.B. and E.P.) independently extracted data according to the prespecified protocol. The procedure of extracting data from publications have been reported elsewhere 23 . Briefly, the renal outcomes of interests were changes in UACR and eGFR, incident microalbuminuria (UACR >30 mg/g) and macroalbuminuria (UACR >300 mg/g), worsening nephropathy (defined as development of microalbuminuria or macroalbuminuria from normoalbuminuria, or progression from microalbuminuria to macroalbuminuria), and the development of ESRD. We extracted mean changes from baseline and their standard deviations for continuous variables in the intervention (SGLT2 inhibitors) and control (placebo or other antidiabetic drugs) groups and used them as the summary measures. The number of patients reporting each renal outcome was extracted for dichotomous variables. We also collected information about the first author, publication year, number and mean age of randomized participants, study duration, intervention and comparison treatment, background therapy for glycemic control, baseline UACR and eGFR, and the history of cardiovascular disease, heart failure, and CKD (eGFR <60 mL/min/1.73 m 2 ). Quality assessment. We used the Cochrane Risk of Bias Tool to assess quality and risk of bias for included studies 24 . Two authors (J.H.B. and E.P.) independently reviewed each RCT and classified the risk of bias as adequate (low risk of bias), unclear (unclear risk of bias), or inadequate (high risk of bias) based on six aspects of trials: sequence generation, allocation concealment, blinding of participants and personnel, incomplete outcome data, selective reporting, and other sources of bias 24  www.nature.com/scientificreports www.nature.com/scientificreports/ Data synthesis and analysis. We calculated weighted mean differences (WMDs) with 95% confidence intervals (CIs) to assess effect size for continuous variables including UACR and eGFR. We also calculated the relative risks (RRs) with 95% CIs to estimate effect size for dichotomous variables including microalbuminuria, macroalbuminuria, worsening nephropathy, and ESRD. In the meta-analysis, we used a random effects model to combine estimators. We also considered a fixed effect model additionally for exploration of the discrepancy in results. The I 2 statistic, τ 2 statistic, and Cochran's Q test were used to assess statistical heterogeneity among the studies 25 . We regarded the I 2 statistic of 0% to 40%, 30% to 60%, 50% to 90%, and 75% to 100% as not significant, moderate, substantial, and considerable heterogeneity, respectively 26 . To detect reporting bias, such as publication bias, asymmetry in the funnel plot was evaluated for renal outcomes 27,28 . We performed a subgroup analysis of eGFR based on baseline eGFR (<60, 60-90, and >90 mL/min/1.73 m 2 ) and study duration (<26, 26-52, and >52 weeks), and sensitivity analyses of microalbuminuria, macroalbuminuria, worsening nephropathy, and ESRD by considering only canagliflozin, dapagliflozin, and empagliflozin for SGLT2 inhibitors. Additionally, prespecified meta-regression was conducted for changes in UACR according to baseline UACR and for changes in eGFR according to baseline eGFR and study duration, respectively. Some outliers were eliminated by diagnostic measures 29 . All statistical analyses were performed using R version 3.1.0 (R Foundation for Statistical Computing, Vienna, Austria). P values of < 0.05 and <0.10 were regarded as statistically significant for treatment effects and test for heterogeneity, respectively.

Results
Characteristics of included studies. The study screening and selection process is shown in Fig. 1. Of 2,421 records retrieved through the database search, 48 studies were included in the systematic review and meta-analysis. Two of the 48 studies 7,30 were pooled analyses of five RCTs 31-35 and another five RCTs 15,36-39 . The characteristics of the included studies and their participants are described in Supplementary Table 1. The total number of participants was 58,165 (34,661 in the SGLT2 inhibitor group and 23,504 in the control group). The number of participants in each study ranged from 114 to 10,142. Three studies had a duration of 187 to 296 weeks 6,9,22 , whereas the remaining studies had a duration ranging from 12 to 104 weeks. The baseline eGFR of the participants was ≥55 (or 60) mL/min/1.73 m 2 in 24 studies 4,8,38-59 , ≥30 mL/min/1.73 m 2 in 14 studies 6,7,9,15,20,22,30,[32][33][34][60][61][62][63] , ≥20 mL/ min/1.73 m 2 and in 1 study 21 . In one study, 74 of 741 participants had an eGFR of ≥15 and <30 mL/min/1.73 m 2 at baseline 35 . Changes in UACR and eGFR. SGLT2 inhibitors significantly lowered the UACR compared with controls (WMD, −14.64 mg/g; 95% CI, −25.15 to −4.12; P = 0.006) (Fig. 2). The test for heterogeneity showed moderate heterogeneity across the studies (I 2 = 53.1%; P = 0.008). In the meta-regression, the UACR-lowering effects of SGLT2 inhibitors tended to be greater with higher levels of baseline UACR (P = 0.081) (Fig. 3A). The changes in eGFR were not significantly different between SGLT2 inhibitors and controls (WMD, 0.19 mL/min/1.73 m 2 ; 95% CI, −0.44 to 0.82; P = 0.552) (Fig. 4A,B). The test for heterogeneity for this showed substantial heterogeneity across the studies (I 2 = 79.6%; P < 0.001). There was a large discrepancy noted in estimated treatment effects between fixed effect and random effects models, depending on weights given to two large trials 20,22 . However, SGLT2 inhibitors significantly slowed the decline in eGFR in patients with >52 weeks of treatment duration compared with controls (Fig. 4B). In the meta-regression, the decline in eGFR were slower in patients with a higher baseline eGFR (P = 0.116) (Fig. 3B) and a longer duration of follow-up (P = 0.038) (Fig. 3C).

Assessment of funnel plot asymmetry.
In the changes in UACR and eGFR, the funnel plots did not show any notable asymmetry apart from a few outlying values ( Supplementary Fig. 3A,B). Although it was hard to determine asymmetry of the plots for in incident microalbuminuria, incident macroalbuminuria, and worsening nephropathy (Supplementary Fig. 4A-C) due to the small number of studies, the funnel plot still appeared to be quite symmetric in the development of ESRD (Supplementary Fig. 4D).

Discussion
This systematic review and meta-analysis found that SGLT2 inhibitors were associated with a significantly lower risk of development or progression of albuminuria compared with placebo or other antidiabetic drugs in patients with type 2 diabetes. The UACR-lowering effects of SGLT2 inhibitors were associated with a higher baseline UACR. The overall changes in eGFR were not different between two groups. However, SGLT2 inhibitors slowed the decline in eGFR in patients with a higher baseline eGFR and a longer duration of treatment. In addition, SGLT2 inhibitor significantly reduced the risk of ESRD compared with controls.
Considering the direct action of SGLT2 inhibitors on the renal tubules and their favorable effects on BP, body weight, and heart failure, these agents have been suggested theoretically to improve renal outcomes in patients with type 2 diabetes 5,66 . The large clinical trials already showed improvement in the composite renal outcomes with SGLT2 inhibitors 6,9,11 . However, these studies were conducted in patients with an average age of 60 years, a long diabetes duration of about 10 years, and an established cardiovascular disease or high cardiovascular risk. In the present study, we demonstrated that SGLT2 inhibitors had the renoprotective effects by reducing the risk of albuminuria and ESRD in patients with a wide range of cardiovascular risk.
SGLT2 inhibitors may reduce albuminuria by several mechanisms including a decrease in glomerular hyperfiltration 67 , improvement in tubulointerstitial fibrosis 68 , systemic BP reduction 69 , changes in plasma volume expansion 70 , and a decrease in uric acid levels 71 . In patients with type 2 diabetes and either microalbuminuria or macroalbuminuria, empagliflozin reduced the UACR independent of changes in hemoglobin A1c (HbA 1c ), BP, and body weight 7 . Dapagliflozin also reduced the UACR for over 2 years of treatment in patients with type 2 diabetes and stage 3 CKD regardless of changes in HbA 1c , BP, eGFR, and uric acid 72 . These findings suggest that SGLT2 inhibitors reduce albuminuria through their direct effects on the kidney. In diabetic mice, SGLT2 inhibitors reduced albuminuria by ameliorating intraglomerular hypertension and tubulointerstitial fibrosis 73,74 , which are the two key contributors to renal damage in diabetic kidney disease (DKD). In line with these findings, our meta-analysis showed that albuminuria-lowering effects of SGLT2 inhibitors were higher on macroalbuminuria than on microalbuminuria. It could be partly explained by the greater UACR reduction in patients with a higher baseline UACR after SGLT2 inhibitor treatment. Therefore, SGLT2 inhibitors may have beneficial effects on albuminuria in the later stage rather than the early stage of DKD, which needs to be evaluated in further studies.
The overall changes in eGFR showed no difference between SGLT2 inhibitors and controls. In the subgroup-analysis and meta-regression, we found that changes in renal function were affected by baseline eGFR and duration of treatment. The changes in renal function after SGLT2 inhibitor treatment were characterized by a rapid decline in eGFR within the first 4-5 weeks, followed by progressive recovery over time 15,35,50 . In addition, the decrease in eGFR was reversible within 2 weeks after drug discontinuation 35 . These findings indicate that the changes in eGFR with SGLT2 inhibitor treatment were a consequence of the drug's hemodynamic effects. In patients with type 1 diabetes, empagliflozin attenuated renal hyperfiltration accompanied by a decrease in eGFR through affecting tubuloglomerular feedback 67 . Reduction in renal hyperfiltration may be beneficial against progressive decline in renal function because intraglomerular hyperfiltration increases the risk of development and progression of DKD [75][76][77] . Besides intrarenal effects, SGLT2 inhibitors may affect eGFR by reducing BP and body weight. In two recent trials, dapagliflozin and empagliflozin similarly maintained BP and weight reduction despite a decline in renal function, whereas HbA 1c reduction was decreased 78,79 , suggesting that they may slow the progressive decline in renal function independent of their glucose-lowering effects 8 .
Finally, SGLT2 inhibitors significantly reduced the risk of ESRD compared with placebo or other antidiabetic drugs. The direction of treatment effects of dapagliflozin was different from those of canagliflozin and empagliflozin but it did not change the overall treatment effect. Considering their beneficial effects on albuminuria, progressive eGFR decline, and glomerular hyperfiltration 4 , SGLT2 inhibitors have been expected to improve hard renal outcomes. In the Canagliflozin and Renal Endpoints in Diabetes with Established Nephropathy Clinical Evaluation (CREDENCE) trial, which evaluated the primary composite renal outcome, canagliflozin showed 30% risk reduction of the composite of ESRD, doubling of serum creatinine, or renal or cardiovascular death, and 32% of risk reduction of ESRD compared with placebo in patients with type 2 diabetes and albuminuric CKD 22 . All the patients in this study were receiving angiotensin-converting-enzyme inhibitor or angiotensin-receptor blocker 22 . Interestingly, the magnitude of renal benefit of SGLT2 inhibitors was greater in patients with less severe kidney disease at baseline in a meta-analysis of cardiovascular outcome trials 80 . Therefore, further investigation is required to determine whether the risk reduction of ESRD is a class effect of SGLT2 inhibitors in patients with or without CKD.
Our study has some limitations. First, most of the studies included in our meta-analysis were originally designed to investigate glucose-lowering effects and safety of SGLT2 inhibitors. Therefore, the analysis of renal outcomes should be interpreted cautiously. Second, 14 of 48 studies reported incomplete outcome data due to losses to follow-up, suggesting the possibility of attrition bias. Third, we could not evaluate the renal effects of SGLT2 inhibitors stratified by cardiovascular risk of included studies because they provided missing or inconsistent data. Recently, another meta-analysis showed that SGLT2 inhibitors reduced the risk of the composite of worsening of renal function, ESRD, or renal death similarly in patients with or without atherosclerotic cardiovascular disease 80 .
In conclusion, our meta-analysis demonstrated that SGLT2 inhibitors had beneficial effects on the kidney by lowering the risk of albuminuria development or progression and reducing the risk of ESRD compared with placebo or other antidiabetic drugs in patients with type 2 diabetes. In addition, the renoprotective effects of SGLT2 inhibitors were greater in patients with a higher UACR and GFR, and a long duration of treatment.