Study characteristics and quality assessment

During the literature search process, studies such as CANVAS 2018 [23] were excluded due to the lack of available data on CKD subgroups with GFR < 60 mL/min/1.73 m2. Eventually, the analysis encompassed 17 RCTs [17, 19, 24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40], incorporating a total of 27,928 participants(Fig. 1).

Fig. 1

Process for identifying studies eligible for the meta-analysis

These studies were conducted from 2014 to 2023, with treatment duration ranging from 18 to 202 weeks and sample sizes ranging from 81 to 10,584(summarized in Table 1). Baseline characteristics in regard to the included RCTs population’s age, sex, CKD stage, level of proteinuria, weight and the percent of diabetes were showed in Table 1.

Table 1 Baseline characteristics of the randomized controlled studies

We evaluated all trials for inherent risk bias using cochrane collaboration tool. As listed in Table S1, the risk of bias across assessed items was generally low in all studies.

Treatment outcomes

Treatment outcomes comprised primary outcomes (cardiovascular events, HHF, renal composite outcomes, cardiovascular death, and all-cause death) and second outcomes(eGFR decline).

Cardiovascular events

Of the seventeen included trials, seven RCTs were analyzed for cardiovascular events. The results indicated that SGLT2i significantly decreased the risks of cardiovascular events (seven studies, 17,355 participants, RR 0.77, 95% CI 0.70–0.84, Fig. 2A) compared to placebo in CKD patients with eGFR < 60 mL/min/1.73 m.

Fig. 2

Effect of SGLT2 inhibitors compared with placebo on cardiovascular outcomes (A), HHF (B) and renal composite outcomes (C)

According to further subgroup analyses, the SGLT2i significantly lowered the risks of cardiovascular outcomes in the subgroup of CKD stage 3a(two studies, 2528 participants, RR 0.81,95% CI 0.72 to 0.93), CKD stage 3b (two studies, 1883 participants, RR 0.76, 95% CI 0.66 to 0.88), CKD stage 3 (four studies, 16,260 participants, RR 0.77, 95% CI 0.72 to 0.82) and CKD stage 4 patients (three studies, 1075 participants, RR 0.76, 95% CI 0.54–0.82), as showed in Fig. 3.

Fig. 3

Summary of subgroup analyses for cardiovascular outcomes of SGLT2 inhibitors compared with control.DKD:diabetic kidney disease;CKD:chronic kidney disease

There was no significant difference in the risk of cardiovascular events between the SGLT2i group and the placebo in the CKD subgroup including non-diabetic patients(one studies, 624 participants, RR 0.76, 95% CI 0.52 to 1.12; p = 0.16, Fig. 3).

The subgroup analyses also indicated that the risks of cardiovascular events were reduced significantly by 28% in the canagliflozin group (two studies, 2766 participants, RR 0.72, 95% CI 0.62 to 0.85), 19% in the dapagliflozin (two studies,1889 participants, RR 0.81, 95% CI 0.68 to 0.96), 16% in the empagliflozin (one studies,1819 participants, RR0.84, 95% CI 0.74 to 0.94) and 28% in the sotagliflozin group(two studies, 10861participants, RR 0.72, 95% CI 0.65 to 0.80) when compared with placebo (Fig. 3).

HHF

Seven RCTs were included for the assessment of HHF. The results indicated that SGLT2i significantly decreased the risks of HHF (seven studies, 17,869 participants, RR 0.73, 95% CI 0.65 to 0.82, Fig. 2B) in CKD patients with eGFR < 60 mL/min/1.73 m2.

As showed in Fig. 4, the subgroup analyses by CKD stages demonstrated that the SGLT2i significantly lowered the risks of HHF in the subgroup of CKD stage 3a by 36%(three studies, 3428 participants, RR 0.64, 95% CI 0.53–0.78), CKD stage 3 by 26% (three studies, 6005 participants, RR 0.74, 95% CI 0.65–0.85) and CKD stage 4 patients by 26% (four studies, 1280 participants, RR 0.74, 95% CI 0.55–1.00). However, it did not achieve significant difference for HHF events between the SGLT2i group and control group for CKD3b patients (three studies, 2577 participants, RR 0.85,95% CI 0.71–1.03, Fig. 4).

Fig. 4

Summary of subgroup analyses for hospitalization for heart failure (HHF) of SGLT2 inhibitors compared with control.DKD:diabetic kidney disease;CKD:chronic kidney disease

We also found that SGLT2i group in CKD subgroup( including non-diabetic patients) also attained significantly decreased risk of HHF (two studies, 2423 participants, RR 0.81, 95% CI 0.70 to 0.94, Fig. 4).

Subgroup analyses by specific SGLT2i agents indicated that the risks of HHF were significantly reduced by 39% in the canagliflozin group (two studies, 2766 participants, RR 0.61, 95% CI 0.45 to 0.82), 22% in the empagliflozin (two studies, 3618 participants, RR 0.78, 95% CI 0.68 to 0.90) and 33% in the sotagliflozin group (two studies, 10,861 participants, RR 0.67, 95% CI 0.52 to 0.85) when compared with placebo (Fig. 4). Dapagliflozin did not show a significant difference compared to the control group (one studies, 624 participants, RR 0.57, 95% CI 0.22 to 1.49, Fig. 4).

Renal composite outcomes

Eight RCTs were included for the assessment of renal composite outcomes. The results indicated that SGLT2i significantly decreased the risks of renal composite outcomes by 30% (eight studies, 22,525 participants, RR 0.70, 95% CI: 0.61 to 0.80; Fig. 2C) compared with placebo in CKD patients with eGFR < 60 mL/min/1.73 m2.

As showed in Fig. 5, subgroup analyses by CKD stages demonstrated that the SGLT2i significantly lowered the risks of renal composite outcomes in CKD stage 3a group by 45%(two studies, 2179 participants, RR 0.55, 95% CI: 0.38–0.78), CKD stage 3b by 25% (three studies, 4935 participants, RR 0.75,95% CI: 0.63 to 0.89), CKD stage 3 by 31% (five studies, 21,891 participants, RR 0.69, 95% CI: 0.61 to 0.80) and CKD stage 4 patients by 22% (six studies, 4375 participants, RR 0.78, 95% CI: 0.68 to 0.88). The study highlighted that renal composite outcomes improved most in patients with CKD stage 3a and least in those with CKD stage 4, though the difference did not reach statistical significance. Additionally, the SGLT2i group in the CKD subgroup including non-diabetic patients also showed significantly lower risk of renal composite outcomes (three studies, 7633 participants, RR 0.75, 95% CI: 0.59–0.94, Fig. 5).

Fig. 5

Summary of subgroup analyses for renal composite outcomes of SGLT2 inhibitors compared with control.DKD:diabetic kidney disease;CKD:chronic kidney disease

Subgroup analyses by specific SGLT2i agents indicated that the risks of renal composite outcomes were significantly reduced by 34% in the canagliflozin group (two studies, 2766 participants, RR 0.66, 95% CI 0.54 to 0.81) and 29% in the dapagliflozin (two studies, 1889 participants, RR 0.71, 95% CI 0.54 to 0.94). However, there was no significant decrease in renal outcomes in the empagliflozin (two studies,7009 participants, RR 0.74, 95% CI 0.53 to 1.02) or sotagliflozin group (two studies,10,861 participants, RR 0.78, 95% CI 0.55 to 1.12) when compared to control group (Fig. 5).

Cardiovascular death

Eight RCTs were included for the assessment of cardiovascular death. The results indicated that SGLT2i significantly decreased the risks of cardiovascular death by 19%(eight studies, 23,079 participants, RR 0.81, 95% CI: 0.74–0.88, Fig. 6A) compared to placebo in CKD patients with eGFR < 60 mL/min/1.73 m2.

Fig. 6

Effect of SGLT2 inhibitors compared with placebo on cardiovascular death (A) and all-cause death (B)

As showed in Fig. 7, subgroup analysis showed a significant reduction in the risk of cardiovascular death in CKD stage 3a by 23% ( three studies, 3428 participants, RR 0.77, 95% CI: 0.63- 0.95) and a 16% decrease in CKD stage 3(five studies, 19,517 participants, RR 0.84, 95% CI: 0.75–0.95). However, no significant difference was observed between the SGLT2i group and control group in CKD stage 3b patients ( four studies, 5505 participants, RR 0.86, 95% CI: 0.71- 1.03) or CKD4 patients( five studies, 3562 participants, RR 0.83, 95% CI: 0.63- 1.10).

Fig. 7

Summary of subgroup analyses for cardiovascular death of SGLT2 inhibitors compared with control.DKD:diabetic kidney disease;CKD:chronic kidney disease

Moreover, the study found that SGLT2i group also attained a significantly reduced risk of cardiovascular death (three studies, 7633 participants, RR 0.81, 95% CI 0.74 to 0.90, Fig. 7) in the CKD subgroup including three RCTs with 3896 (51%) non-diabetic patients.

Subgroup analyses also indicated that the risks of cardiovascular death were significantly reduced by 19% in the empagliflozin group (three studies, 8828 participants, RR 0.81, 95% CI 0.73 to 0.89). It did not achieve significant difference in cardiovascular death in the canagliflozin (two studies, 2766 participants, RR 0.76, 95% CI 0.58 to 1.01), sotagliflozin (two studies, 10,861 participants, RR 0.81, 95% CI 0.62 to 1.10) and dapagliflozin group (one study, 624 participants, RR 0.81, 95% CI 0.36 to 1.79) when compared to control group (Fig. 7).

All-cause death

Five RCTs were included for the assessment of all-cause death. There was no statistical difference in the risk of all-cause death between the SGLT2i and control groups (five studies, 13,724 participants, RR 0.92, 95% CI 0.80 to 1.06, Fig. 6B). Subgroup analyses for all-cause death were not conducted due to a lack of available data.

The decline rate of eGFR

The decline in eGFR was assessed through total slopes(A) and chronic slopes(B) between SGLT2i and placebo group (Fig. 8). SGLT2i significantly slowed eGFR decline: total slopes (five studies, 10,370 participants, mean difference 1.17, 95%CI 0.86–1.49) and chronic slopes (four studies, 8459 participants, mean difference 2.12, 95%CI 1.64–2.61).

Fig. 8

The mean rate of eGFR decline with SGLT2i compared with placebo: Total slopes (A) and Chronic slopes (B)

Subgroup analyses indicated that SGLT2i mitigated the eGFR decline in both CKD stage 3 patients (total slopes: three studies, 7085 participants, mean difference 1.24, 95%CI 0.65–1.83;chronic slopes: two studies, 5375 participants, mean difference 2.29, 95%CI 1.25–3.33, Table 2)

Table 2 Subgroup analysis for the mean rate of eGFR decline with SGLT2i compared with placebo: Total slopes and Chronic slopes

and CKD stage 4 patients (total slopes: four studies, 3285 participants, mean difference 0.91, 95%CI 0.87–0.95;chronic slopes: three studies, 3080 participants, mean difference 1.96, 95%CI 0.88–3.05, Table 2).

Subgroup analyses highlighted that canagliflozin was the most effective in slowing eGFR decline (total slopes: two studies, 2737 participants, mean difference 1.77, 95%CI 1.65–1.89;chronic slopes: two studies, 2625 participants, mean difference 2.64, 95%CI 2.51–2.77, Table 2), followed by dapagliflozin (total slopes: one study, 624 participants, mean difference 1.23, 95%CI 1.18–1.28; chronic slopes: one study, 624 participants, mean difference 2.35, 95%CI 2.31–2.39, Table 2) and empagliflozin (total slopes: two studies, 7459 participants, mean difference 0.80, 95%CI 0.61–0.99; chronic slopes: one study, 5210 participants, mean difference 1.17, 95%CI 0.86–1.47, Table 2), and significant inter-subgroup variations were observed for total slopes(p < 0.001) and chronic slopes(p < 0.001, Table 2) among canagliflozin, dapagliflozin, and empagliflozin groups compared to placebo.

Subgroup analyses also demonstrated that SGLT2i most effectively decelerated eGFR decline in diabetes patients with chronic kidney disease (total slopes: two studies, 2737 participants, mean difference 1.77, 95%CI 1.65–1.89; chronic slopes: two studies, 2625 participants, mean difference 2.64, 95%CI 2.51–2.77, Table 2), followed in chronic kidney disease(CKD) patients(total slopes: three studies, 7633 participants, mean difference 0.87, 95%CI 0.69–1.06; chronic slopes: two studies, 5834 participants, mean difference 1.56, 95%CI 1.16–1.96, Table 2) compared to placebo. Notable differences between diabetic kidney disease and CKD populations were evident for both total slopes(p < 0.001, Table 2) and chronic slopes(p < 0.001, Table 2).

Subgroup analyses also indicated significantly slower reduction in eGFR was achieved with longer exposure to SGLT2i (duration > 100weeks, total slopes: four studies, 8571 participants, mean difference 1.30, 95%CI 0.92–1.69,) compared to placebo. However, no significant distinction was found for duration less than 100 weeks (total slopes: one study, 1799 participants, mean difference 0.90, 95%CI -0.03–1.83, Table 2). Regrettably, there was no significant difference between subgroups based on duration above and below 100 weeks(p = 0.43, Table 2).

Adverse effects

Table 3 showed summarized clinical trial results for adverse events. The SGLT2i group exhibited statistically lower risks of (fourteen studies, 19,654 participants, RR 0.91, 95% CI 0.87–0.95), hypoglycemia(nine studies, 16,412 participants, RR 0.91, 95% CI 0.84–0.98), hyperkalemia (four studies, 2693 participants, RR 0.68, 95% CI 0.51–0.93) and acute renal injury(five studies, 5424 participants, RR 0.79, 95% CI 0.65–0.95) compare to placebo group. No increased risk of any adverse event (fourteen studies, 19,674 participants, RR 0.99, 95% CI 0.99 to 1.01), any adverse effects leading to discontinuation(eleven studies,16,638 participants, RR 0.95, 95% CI 0.80 to 1.12), fracture(nine studies, 18,493 participants, RR 0.93, 95% CI 0.79 to 1.10), urinary tract infection(ten studies, 15,987 participants, RR 1.05, 95% CI 0.97 to 1.15) or amputation(three studies, 13,800 participants, RR 0.95, 95% CI 0.69 to 1.31) was associated with SGLT2i use. However, SGLT2i increased the morbidity rate of genital infection(eleven studies, 16,284 participants, RR 2.70, 95% CI 2.09 to 3.49) and volume depletion (thirteen studies, 20,664 participants, RR 1.20, 95% CI 1.04 to 1.39) in CKD patients with eGFR < 60 mL/min/1.73 m2.

Table 3 Comparison of adverse events between the SGLT2 inhibitors and placebo group in patients with eGFR < 60 mL/min/1.73 m2 and eGFR < 30 mL/min/1.73 m2

Further subgroup analyses indicated that no significant differences in the risks of any adverse effects (three studies, 525 participants, RR 1.03, 95% CI 0.96 to 1.10), any serious adverse effects(four studies, 1149 participants, RR 0.88, 95% CI 0.75 to 1.04), genital infection(two studies, 349 participants, RR2.18 95% CI 0.23 to 20.63), volume depletion(three studies, 975 participants, RR 1.01, 95% CI 0.60 to 1.80), hypoglycemia(three studies, 975 participants, 0.94, 95% CI 0.62 to 1.43) or hyperkalemia(two studies, 248 participants, RR0.64, 95% CI 0.11 to 3.72) between the two treatment groups in the subgroup of CKD stage 4 patients (Table 3).

Sensitivity analysis and publication bias

The heterogeneity was generally low with I2 < 40% for all treatment outcomes and safety outcomes in CKD patients with eGFR < 60 mL/min/1.73 m2. We conducted sensitivity analyses for various outcomes, including cardiovascular events, HHF, renal composite outcomes, cardiovascular death, all-cause death, any adverse event, any serious adverse events and hyperkalemia. The results showed that the study EMPEROR-Reduced 2021 had a great impact on the final result of HHF in CKD patients with eGFR < 60 mL/min/1.73 m2 (Fig S1B). Besides, the study SCORED2021 had a significant impact on the final result of all- cause death (Fig S3A). No studies had a significant impact on the rest of final outcomes of cardiovascular events(Fig S1A), cardiovascular death(Fig S2A), renal composite outcomes(Fig S2B), any adverse event(Fig S3B), any serious adverse event(Fig S4A) and hyperkalemia (Fig S4B) in CKD patients with eGFR < 60 mL/min/1.73 m2.

Statistical results using Egger's test showed there was no publication bias for cardiovascular events (7 studies, p = 0.526), HHF(7 studies, p = 0.224), renal composite outcomes(8 studies, p = 0.528), cardiovascular death(5 studies, p = 0.394), all-cause death(5 studies, p = 0.235), any adverse event(14 studies,p = 0.236) and any serious adverse event(14 studies, p = 0.972) in CKD patients with eGFR < 60 mL/min/1.73 m2(Table S2).

Egger's test showed there was no publication bias for cardiovascular events (3 studies, p = 0.454), HHF(4 studies, p = 0.458), renal composite outcomes(6 studies, p = 0.054), cardiovascular death(5 studies, p = 0.275),any adverse event(3 studies, p = 0.282) and any serious adverse event(4 studies, p = 0.173) in CKD patients with eGFR < 30 mL/min/1.73 m2(Table S2).

The quality of evidence

The quality of evidence was assessed using the GRADE (Grading of Recommendations Assessment, Development and Evaluation) system. The summary of findings presented in supplementary appendix(Table S3 and Table S4). Table S3 showed the results of GRADE Assessment for patients with eGFR < 60 mL/min/1.73 m2 and Table S4 was for patients with eGFR < 30 mL/min/1.73 m2. The quality of evidence for most of the SGLT2i-related outcome indicators in patients with eGFR < 60 mL/min/1.73 m2 was high. However, the evidence for outcome indicators related to cardiovascular outcomes, genital infection and fracture was moderate due to the high risk of inconsistent study results. For patients with eGFR < 30 mL/min/1.73 m2, the evidence for outcome indicators related to hyperkalemia was low and the quality for cardiovascular outcomes, cardiovascular death, genital infection and hypoglycemia was moderate due to the high risk of inconsistency and imprecision. For the rest of outcome indicators in patients with eGFR < 30 mL/min/1.73 m2, the evidence was in high level.