Literature search and included studies
The detailed study filtering process is shown in Fig 1. In brief, we retrieved a total of 5163 articles from PubMed(n=977), Cochrane Library(n=74), Web of science(n=1022), Medline(n=1470) and Embase(n=1620) in primary search, during the processanother 12 articles were identified through references. A total of 2232 duplicate articles were removed. After review by title and abstract, 94 articles remained and entered into full-text assessing section. By assessing full text, 56 additional articles with missing data were excluded. Finally, 38 articles (included 17 randomized controlled trials) were included in this network meta-analysis [7,16,17,28-62]. Out of 17 studies, 3 studies were compared finerenone [16,17,28-30] with placebo; 8 studies were compared SGLT2i (Empaglifozin [31-35], Canaglifozin [7,36-42], Dapaglifozin [43-47], Ertuglifozin [48-50], and Sotaglifozin [51,52]) with placebo; 6 studies compared GLP-1 RA (Dulaglutide [53,54], Albiglutide [55], Exenatide [56,57], Semaglutide [58,59] and Liraglutide [60-62]) with placebo.
Baseline characteristics of included studies in patients with T2DM and CKD
The characteristics of the included studies are presented in Table 1. The pooled population consisted of 51,496 patients with T2DM and CKD, 14,847 of them were in finerenone studies (7246 in the intervention group and 7601 in control group), 25,098 patients in SGLT-2i studies (13,260 in the intervention group and 11,838 in control group) and 11,551 patients in GLP-1 RA studies (5355 in the group treated with GLP-1 RA and 5796 in the control group). The definition of MACE in the included trials was consistent, except for three of them, EMPA-REG, DECLARE–TIMI 58 and EXSCEL trials (data for nonfatal MI and stroke were not available, so we used total MI and stroke instead). Whereas renal events were defined slightly different across included trials, but they were similar enough that can be used in analysis. The detailed definitions of renal events in different trials are shown in Table 2.
Table 2
Definitions of terms in included studies
Trial
|
Study design
|
Patients enrolled in trials
|
Patients included in this study
|
Setting
|
Drug dose (mg/day)
|
Median follow up
|
eGFR
|
Range of HbA1c (%)
|
Definitions of renal events among included trials in patients with T2DM and CKD
|
Finerenone vs placebo
|
FIDELIO-DKD
|
RCT
|
T2DM and CKD
|
T2DM and CKD
|
Multinational
|
Finerenone 10/20
|
2.6years
|
25 to <75
|
≤12
|
≥40% eGFR decline, renal death, ESRD, eGFR<15 ml/min/1.73 m2
|
FIGARO-DKD
|
RCT
|
T2DM and CKD
|
T2DM and CKD
|
Multinational
|
Finerenone 10/20
|
3.4years
|
25 to 90
|
≤12
|
≥40% eGFR decline, renal death
|
ARTS-DN
|
RCT
|
DN
|
DN
|
Multinational
|
Finerenone 1.25/2.5/5/7.5/10/15/20
|
90 days
|
≥30
|
≤12
|
≥40% eGFR decline
|
SGLT2i vs placebo
|
EMPA-REG
|
RCT
|
T2DM
|
T2DM and CKD
|
Multinational
|
Empagliflozin 10/25
|
3.1 years
|
≥30
|
7 to 10
|
Macroalbuminuria, doubling of serum creatinine, eGFR<45 ml/min/1.73 m2, renal-replacement therapy; renal death
|
CANVAS
|
RCT
|
T2DM
|
T2DM and CKD
|
Multinational
|
Canagliflozin 100/300
|
188.2weeks
|
≥30
|
7 to 10.5
|
ESRD, renal death, ≥40% eGFR decline, doubling of serum creatinine.
|
DECLARE–TIMI 58
|
RCT
|
T2DM
|
T2DM and CKD
|
Multinational
|
Dapagliflozin 10
|
4.2 years
|
CrCl≥60ml/min
|
6.5 to 12
|
≥40% eGFR decline, renal death, ESRD
|
CREDENCE
|
RCT
|
T2DM and CKD
|
T2DM and CKD
|
Multinational
|
Canagliflozin 100
|
2.62years
|
30 to <90
|
6.5 to 12
|
ESRD, doubling of serum creatinine level, renal death
|
VERTIS CV
|
RCT
|
T2DM
|
T2DM and CKD
|
Multinational
|
Ertugliflozin 5/15
|
3.5years
|
≥30
|
7 to 10.5
|
N/A
|
DAPA-CKD
|
RCT
|
CKD
|
T2DM and CKD
|
Multinational
|
Dapagliflozin 10
|
2.4 years
|
25 to 75
|
N/A
|
≥50% eGFR decline, ESRD, renal-replacement therapy, eGFR<15 ml/min/1.73 m2, renal death
|
SCORED
|
RCT
|
T2DM and CKD
|
T2DM and CKD
|
Multinational
|
Sotagliflozin 400
|
16 months
|
25 to 60
|
≥7
|
≥50% eGFR decline, renal-replacement therapy, eGFR<15 ml/min/1.73 m2
|
Cherney 2021
|
RCT
|
T2DM and CKD
|
T2DM and CKD
|
Multinational
|
Sotagliflozin 200/400
|
52weeks
|
15 to <30
|
7 to 11
|
≥50% eGFR decline, renal-replacement therapy, eGFR<15 ml/min/1.73 m2, renal death
|
GLP-1 RA vs placebo
|
LEADER
|
RCT
|
T2DM
|
T2DM and CKD
|
Multinational
|
Liraglutide 1.8
|
3.84years
|
N/A
|
≥7
|
Macroalbuminuria, doubling of serum creatinine, eGFR<45 ml/min/1.73 m2, renal-replacement therapy, renal death
|
REWIND
|
RCT
|
T2DM
|
T2DM and CKD
|
Multinational
|
dulaglutide 1.5 weekly
|
5.4years
|
≥15
|
≤9.5
|
Macroalbuminuria, ≥30% eGFR decline, renal-replacement therapy,
|
HARMONY
|
RCT
|
T2DM
|
T2DM and CKD
|
Multinational
|
Albiglutide 30/50
|
1.5years
|
≥30
|
>7
|
N/A
|
EXSCEL
|
RCT
|
T2DM
|
T2DM and CKD
|
Multinational
|
Exenatide 2 weekly
|
3.2years
|
≥30
|
6.5 to 10
|
≥40% eGFR decline, renal-replacement therapy, renal death
|
PIONEER-6
|
RCT
|
T2DM
|
T2DM and CKD
|
Multinational
|
Semaglutide 14 oral
|
15.9months
|
≥30
|
N/A
|
N/A
|
SUSTAIN-6
|
RCT
|
T2DM
|
T2DM and CKD
|
Multinational
|
Semaglutide 0.5/1 weekly
|
109weeks
|
N/A
|
≥7
|
N/A
|
Note: DN: Diabetic Nephropathy; N/A: not available; CrCl: creatinine clearance; eGFR (ml/min/1.73m2)
|
Risk of bias
We assess the risk of bias in those trials using the Revised Cochrane Risk of Bias Tool (RoB 2.0). The quality evaluation of the included studies is shown in Fig 2. All trials were evaluated as low risk in 5 outcomes. Detailed evaluations are as shown in Additional file 1 (RoB-2 evaluation) for each study.
GRADE assessment
In terms of reducing the MACE, there were 14 direct comparisons in the original articles and they were estimated high quality. In terms of reducing the renal outcomes, there were 13 direct comparisons in the original articles whose estimated results were high quality. In terms of reducing the HHF and CVD, there were 12 direct comparisons in the original articles and they were rated as high quality. In terms of reducing the ACD, there were 11 direct comparisons in the original articles and they were rated as high quality. The detail was shown in Table 3. Fig 3 shows the network graph.
According to recommendation of GRADE working group, we presented a four-step approach to rate the quality of evidence in each of the direct, indirect, and network meta-analysis estimates based on methods developed by the GRADE working group [26]. In this network meta-analysis, none of the 5 outcomes had a closed loop. Meaning that that no outcomes from both direct and indirect comparisons are included. Thus, rendering incoherence assessment unnecessary. The studies that are included had slightly different definitions of renal outcomes, and the baseline eGFR of patients in the “cherney 2021” was different from other studies. For direct comparisons, “cherney 2021” included only 1% of patients in SGLT2i (277/25098). Therefore, risk of bias was not taken in to consideration. As for intransitivity, there was only indirect evidence in the intercomparison of the three drugs. The GRADE working group recommends in such situation issues regarding intransitivity may warrant particular attention, and the threshold for rating down for intransitivity may be lower [26]. Therefore, we downgraded the quality of evidence for the comparison between SGLT2i and the two other drugs. The detail was shown in Table 4.
Table 3
GRADE assessment
|
Certainty assessment
|
№ of patients
|
Certainty
|
Importance
|
Intervention of studies
|
Study design
|
Risk of
bias
|
Inconsistency
|
Indirectness
|
Imprecision
|
Publication
bias
|
Intervention group
|
Control
group
|
MACE (№ of studies: 14) CRITICAL
|
SGLT2i vs Placebo
|
randomised trials
|
not serious
|
not serious
|
not serious
|
not serious
|
none
|
1255/13143 (9.5%)
|
1266/11688 (10.8%)
|
High
|
|
GLP-1 RA vs Placebo
|
randomised trials
|
not serious
|
not serious
|
not serious
|
not serious
|
none
|
633/4674 (13.5%)
|
721/4678 (15.4%)
|
High
|
|
Finerenone vs Placebo
|
randomised trials
|
not serious
|
not serious
|
not serious
|
not serious
|
none
|
228/2833 (8.0%)
|
258/2841 (9.1%)
|
High
|
|
Renal outcomes (№ of studies: 13)
|
|
|
|
|
|
|
CRITICAL
|
SGLT2i vs Placebo
|
randomised trials
|
not serious
|
not serious
|
not serious
|
not serious
|
none
|
586/11731 (5.0%)
|
700/10981 (6.4%)
|
High
|
|
GLP-1 RA vs Placebo
|
randomised trials
|
not serious
|
not serious
|
not serious
|
not serious
|
none
|
428/3754 (11.4%)
|
474/3780 (12.5%)
|
High
|
|
Finerenone vs Placebo
|
randomised trials
|
not serious
|
not serious
|
not serious
|
not serious
|
none
|
861/7234 (11.9%)
|
997/6600 (15.1%)
|
High
|
|
HHF (№ of studies: 12)
|
|
|
|
|
|
|
|
IMPORTANT
|
SGLT2i vs Placebo
|
randomised trials
|
not serious
|
not serious
|
not serious
|
not serious
|
none
|
395/13144 (3.0%)
|
561/11689 (4.8%)
|
High
|
|
GLP-1 RA vs Placebo
|
randomised trials
|
not serious
|
not serious
|
not serious
|
not serious
|
none
|
174/2673 (6.5%)
|
191/2662 (7.2%)
|
High
|
|
Finerenone vs Placebo
|
randomised trials
|
not serious
|
not serious
|
not serious
|
not serious
|
none
|
256/6519 (3.9%)
|
325/6507 (5.0%)
|
High
|
|
ACD (№ of studies: 11) IMPORTANT
|
SGLT2i vs Placebo
|
randomised trials
|
not serious
|
not serious
|
not serious
|
not serious
|
none
|
706/12128 (5.8%)
|
720/10898 (6.6%)
|
High
|
|
GLP-1 RA vs Placebo
|
randomised trials
|
not serious
|
not serious
|
not serious
|
not serious
|
none
|
339/2673 (12.7%)
|
378/2662 (14.2%)
|
High
|
|
Finerenone vs Placebo
|
randomised trials
|
not serious
|
not serious
|
not serious
|
not serious
|
none
|
552/6519 (8.5%)
|
614/6507 (9.4%)
|
High
|
|
CVD (№ of studies:12) IMPORTANT
|
SGLT2i vs Placebo
|
randomised trials
|
not serious
|
not serious
|
not serious
|
not serious
|
none
|
623/13144 (4.7%)
|
597/11689 (5.1%)
|
High
|
|
GLP-1 RA vs Placebo
|
randomised trials
|
not serious
|
not serious
|
not serious
|
not serious
|
none
|
213/2673 (8.0%)
|
235/2662 (8.8%)
|
High
|
|
Finerenone vs Placebo
|
randomised trials
|
not serious
|
not serious
|
not serious
|
not serious
|
none
|
322/6519 (4.9%)
|
364/6507 (6.0%)
|
High
|
|
Table 4
Estimates of effects and quality ratings for comparison of drugs to prevent cardiorenal outcomes
|
Comparison
|
Direct evidence
|
Indirect evidence
|
Network meta-analysis
|
RR [95% CI]
|
Quality of evidence
|
RR [95% CI]
|
Quality of evidence
|
RR [95% CI]
|
Quality of evidence
|
MACE
|
|
|
|
|
|
|
SGLT2i vs Placebo
|
0.83 (0.77,0.90)
|
High
|
Not estimable▲
|
—
|
0.84 (0.78,0.90)
|
High
|
GLP-1 RA vs Placebo
|
0.88 (0.80,0.97)
|
High
|
Not estimable▲
|
—
|
0.88 (0.80,0.97)
|
High
|
Finerenone vs Placebo
|
0.89 (0.75,1.05)
|
High
|
Not estimable▲
|
—
|
0.89 (0.75,1.05)
|
High
|
GLP-1 RA vs SGLT2i
|
—
|
—
|
1.05 (0.93,1.19)
|
Moderate*
|
1.05 (0.93,1.19)
|
Moderate*
|
Finerenone vs SGLT2i
|
—
|
—
|
1.06 (0.88,1.28)
|
Moderate*
|
1.06 (0.88,1.28)
|
Moderate*
|
Finerenone vs GLP-1 RA
|
—
|
—
|
1.01 (0.83,1.23)
|
High
|
1.01 (0.83,1.23)
|
High
|
Renal outcomes
|
|
|
|
|
|
|
SGLT2i vs Placebo
|
0.67 (0.60,0.74)
|
High
|
Not estimable▲
|
—
|
0.67 (0.60,0.74)
|
High
|
GLP-1 RA vs Placebo
|
0.90 (0.80,1.02)
|
High
|
Not estimable▲
|
—
|
0.90 (0.80,1.02)
|
High
|
Finerenone vs Placebo
|
0.86 (0.79,0.93)
|
High
|
Not estimable▲
|
—
|
0.86 (0.79,0.93)
|
High
|
GLP-1 RA vs SGLT2i
|
—
|
—
|
1.36 (1.16,1.59)
|
Moderate*
|
1.36 (1.16,1.59)
|
Moderate*
|
Finerenone vs SGLT2i
|
—
|
—
|
1.29 (1.13,1.47)
|
Moderate*
|
1.29 (1.13,1.47)
|
Moderate*
|
Finerenone vs GLP-1 RA
|
—
|
—
|
0.95 (0.82,1.10)
|
High
|
0.95 (0.82,1.10)
|
High
|
HHF
|
|
|
|
|
|
|
SGLT2i vs Placebo
|
0.60 (0.53,0.68)
|
High
|
Not estimable▲
|
—
|
0.60 (0.53,0.68)
|
High
|
GLP-1 RA vs Placebo
|
0.90 (0.74,1.09)
|
High
|
Not estimable▲
|
—
|
0.90 (0.73,1.09)
|
High
|
Finerenone vs Placebo
|
0.79 (0.67,0.92)
|
High
|
Not estimable▲
|
—
|
0.79 (0.67,0.92)
|
High
|
GLP-1 RA vs SGLT2i
|
—
|
—
|
1.49 (1.18,1.89)
|
Moderate*
|
1.49 (1.18,1.89)
|
Moderate*
|
Finerenone vs SGLT2i
|
—
|
—
|
1.31 (1.07,1.61)
|
Moderate*
|
1.31 (1.07,1.61)
|
Moderate*
|
Finerenone vs GLP-1 RA
|
—
|
—
|
0.88 (0.68,1.14)
|
High
|
0.88 (0.68,1.14)
|
High
|
CVD
|
|
|
|
|
|
|
SGLT2i vs Placebo
|
0.86 (0.77,0.96)
|
High
|
Not estimable▲
|
—
|
0.86 (0.77,0.96)
|
High
|
GLP-1 RA vs Placebo
|
0.90 (0.75,1.08)
|
High
|
Not estimable▲
|
—
|
0.90 (0.75,1.08)
|
High
|
Finerenone vs Placebo
|
0.88 (0.76,1.02)
|
High
|
Not estimable▲
|
—
|
0.88 (0.76,1.02)
|
High
|
GLP-1 RA vs SGLT2i
|
—
|
—
|
1.04 (0.85,1.29)
|
Moderate*
|
1.04 (0.85,1.29)
|
Moderate*
|
Finerenone vs SGLT2i
|
—
|
—
|
1.02 (0.85,1.23)
|
Moderate*
|
1.02 (0.85,1.23)
|
Moderate*
|
Finerenone vs GLP-1 RA
|
—
|
—
|
0.98 (0.78,1.23)
|
High
|
0.98 (0.78,1.23)
|
High
|
ACD
|
|
|
|
|
|
|
SGLT2i vs Placebo
|
0.90 (0.81,0.99)
|
High
|
Not estimable▲
|
—
|
0.89 (0.81,0.99)
|
High
|
GLP-1 RA vs Placebo
|
0.89 (0.78,1.02)
|
High
|
Not estimable▲
|
—
|
0.89 (0.77,1.02)
|
High
|
Finerenone vs Placebo
|
0.90 (0.80,1.00)
|
High
|
Not estimable▲
|
—
|
0.90 (0.80,1.00)
|
High
|
GLP-1 RA vs SGLT2i
|
—
|
—
|
0.99 (0.84,1.18)
|
Moderate*
|
0.99 (0.84,1.18)
|
Moderate*
|
Finerenone vs SGLT2i
|
—
|
—
|
1.00 (0.86,1.16)
|
Moderate*
|
1.00 (0.86,1.16)
|
Moderate*
|
Finerenone vs GLP-1 RA
|
—
|
—
|
1.01 (0.85,1.20)
|
High
|
1.01 (0.85,1.20)
|
High
|
▲Cannot be estimated because the drug was not connected in a loop in the evidence network *Intransitivity
|
Network meta‑analysis of treatment groups
MACE
The results showed that in patients with T2DM and CKD, compared with placebo, SGLT-2i (RR [95% CI]; 0.84 [0.78–0.90]) and GLP-1 RA (RR [95% CI]; 0.88 [0.80–0.97]) were associated with a decreased risk of MACE, while finerenone (RR [95% CI]; 0.89 [0.75–1.05]) did not. Finerenone didn`t show a significant difference in the risk of MACE compared with SGLT-2i (RR [95% CI]; 1.19 [0.92–1.53]) and GLP-1 RA (RR [95% CI]; 1.02 [0.78–1.32]). There was also no significant difference in the risk of MACE between SGLT-2i and GLP-1 RA (RR [95% CI]; 1.17 [0.93–1.47]). There was no heterogeneity (I2=3.3%, p=0.409). The detail is shown in Fig.4.
Renal outcome
Finerenone (RR [95% CI]; 0.86 [0.79–0.93]) and SGLT-2i (RR [95% CI]; 0.67 [0.60–0.74]) significantly decreased the morbidity of renal events, while GLP-1 RA (RR [95% CI]; 0.90 [0.73–1.02]) did not. Compared with finerenone (RR [95% CI]; 1.31 [1.07–1.61]) and GLP-1 RA (RR [95% CI]; 1.49 [1.18–1.89]), SGLT-2i were associated with a decreased morbidity of renal events. Finerenone was comparable to GLP-1 RA (RR [95% CI]; 0.95 [0.82–1.10]). There was moderate heterogeneity (I2=37.4%, p=0.085). The detail is shown in Fig.5.
HHF
Compared with placebo, finerenone (RR [95% CI]; 0.79 [0.67–0.92]) was associated with a decreased risk of HHF while GLP-1 RA (RR [95% CI]; 0.90 [0.73–1.09]) did not. But there was no significant difference in the risk of HHF between the two drugs (RR [95% CI]; 0.88 [0.68–1.14]). Compared with finerenone (RR [95% CI]; 1.31 [1.07–1.61]), GLP-1 RA (RR [95% CI]; 1.49 [1.18–1.89]) and placebo (RR [95% CI]; 0.60 [0.53–0.68]), SGLT-2i were shown to be significantly more effective in reducing HHF. There was moderate heterogeneity (I2=44.9%, p=0.046). The detail is shown in Fig.6.
ACD
Finerenone (RR [95% CI]; 0.90 [0.80–1.00) had a tendency to decrease the risk of ACD and SGLT-2i (RR [95% CI]; 0.89 [0.81–0.99]) were associated with a decreased risk of ACD, while GLP-1 RA (RR [95% CI]; 0.89 [0.77–1.02]) did not. And there was no significant difference among the three drugs (RR 0.99, 95% CI 0.84–1.18; RR 1.00, 95% CI 0.86–1.16; RR 1.01, 95% CI 0.85–1.20, respectively). This analysis showed no heterogeneity (I2=0.0%, p=0.554). The detail is shown in Fig.7.
CVD
As for CVD, compared with placebo, only SGLT-2i were associated with a decreased events (RR [95% CI]; 0.86, [0.77–0.96]). There was no significant difference between the other two drugs and placebo, or among the three drugs (Fig.8). The analysis of CVD showed no heterogeneity (I2=4.4%, p=0.402). The detail is shown in Fig.8.
Specific drugs’ Pairwise comparison and analysis
In order to provide more specific recommendations for clinical drug selection, we further evaluated the efficacy of specific drugs based on its component. Compared to placebo, canagliflozin(RR [95% CI]; 0.78 [0.68–0.89]) , sotagliflozin(RR [95% CI]; 0.76 [0.66–0.87]) and liraglutide(RR [95% CI]; 0.69 [0.58–0.82]) displayed significantly lower risk of MACE,while finerenone (RR [95% CI]; 0.89 [0.75–1.05]), empagliflozin(RR [95% CI]; 0.89 [0.71–1.12]), dapagliflozin(RR [95% CI]; 0.92 [0.77–1.11]), ertugliflozin(RR [95% CI]; 1.09 [0.87–1.36]), , albiglutide(RR [95% CI]; 0.93 [0.73–1.18]), exenatide(RR [95% CI]; 1.03 [0.89–1.20]) and semaglutide(RR [95% CI]; 0.83 [0.60–1.13])were not. Compared to ertugliflozin or exenatide, canagliflozin, sotagliflozin and liraglutide were associated with a decreased risk of MACE. Liraglutide had a tendency to reduce MACE compared to finerenone (RR [95% CI]; 0.78 [0.61–1.00]) and albiglutide (RR [95% CI]; 0.74 [0.55–1.00]), it also showed more positive influence when compared with dapagliflozin (RR [95% CI]; 0.75 [0.58–0.96]). The detail is shown in Table 5a.
In renal outcome, finerenone (RR [95% CI]; 0.86 [0.79–0.93]), empagliflozin (RR [95% CI]; 1.55 [1.01–2.39]), canagliflozin (RR [95% CI]; 1.55 [1.01–2.39]) and dapagliflozin (RR [95% CI]; 1.55 [1.01–2.39]) reduced renal events significantly compared to placebo. The results of comparison showed that empagliflozin (RR [95% CI]; 0.76 [0.63–0.93]), canagliflozin (RR [95% CI]; 0.81 [0.67–0.99]) and dapagliflozin (RR [95% CI]; 0.70 [0.55–0.87]) significantly reduced the morbidity of renal events compared to finerenone. The detail is shown in Table 5b.
Finerenone(RR [95% CI]; 0.79 [0.67–0.92]), empagliflozin(RR [95% CI]; 0.59 [0.40–0.88]), canagliflozin(RR [95% CI]; 0.60 [0.48–0.75]), dapagliflozin(RR [95% CI]; 0.56 [0.41–0.76]), ertugliflozin(RR [95% CI]; 0.51 [0.34–0.76]), sotagliflozin(RR [95% CI]; 0.67 [0.52–0.85]) and liraglutide(RR [95% CI]; 0.73 [0.56–0.97]) reduced HHF significantly, but exenatide(RR [95% CI]; 1.10 [0.83–1.46]) did not. At the same time, all of the 7 drugs mentioned above significantly reduced HHF compared to exenatide (Table 5a). Another discovery worth noting is that canagliflozin (RR [95% CI]; 0.76 [0.58–1.00]) and dapagliflozin (RR [95% CI]; 0.71 [0.50–1.00]) had a tendency to decrease HHF compared to finerenone, and finerenone was associated with a higher risk of HHF than ertugliflozin (RR [95% CI]; 1.55 [1.01–2.39]). The detail is shown in Table 5a.
When it comes to ACD, dapagliflozin (RR [95% CI]; 0.81 [0.66–0.98]) and liraglutide (RR [95% CI]; 0.76 [0.62–0.93]) had significant effect than placebo. And finerenone (RR [95% CI]; 0.90 [0.80–1.00]) tended to reduce the risk of ACD when compared with placebo. As for CVD, liraglutide (RR [95% CI]; 0.69 [0.52–0.90]) was better than placebo, while other drugs were not. The detail is shown in Table 5c.
Table 5a
Pairwise league table of MACE and HHF.
Table 5a - Comparisons for MACE (bottom left) of the 10 drugs and HHF (upper right) of the 8 drugs. RR with 95%CI
|
EM
|
1.01 (0.65,1.59)
|
0.94 (0.57,1.55)
|
0.85 (0.49,1.50)
|
1.12 (0.71,1.78)
|
1.33 (0.87,2.03)
|
|
1.85 (1.14,3.01)
|
|
1.24 (0.76,2.00)
|
1.68 (1.14,2.50)
|
1.14 (0.88,1.49)
|
CA
|
0.93 (0.64,1.35)
|
0.84 (0.53,1.33)
|
1.11 (0.80,1.53)
|
1.31 (1.00,1.71)
|
|
1.83 (1.28,2.61)
|
|
1.22 (0.86,1.74)
|
1.66 (1.34,2.06)
|
0.96 (0.72,1.29)
|
0.84 (0.67,1.06)
|
DA
|
0.91 (0.55,1.51)
|
1.19 (0.81,1.76)
|
1.41 (1.00,2.00)
|
|
1.97 (1.30,2.99)
|
|
1.31 (0.87,1.99)
|
1.79 (1.32,2.44)
|
0.81 (0.59,1.13)
|
0.71 (0.54,0.93)
|
0.84 (0.63,1.13)
|
ER
|
1.32 (0.82,2.10)
|
1.55 (1.01,2.39)
|
|
2.17 (1.33,3.54)
|
|
1.45 (0.89,2.36)
|
1.97 (1.32,2.95)
|
1.17 (0.90,1.53)
|
1.02 (0.85,1.24)
|
1.22 (0.97,1.52)
|
1.44 (1.10,1.89)
|
SO
|
1.18 (0.89,1.58)
|
|
1.65 (1.14,2.39)
|
|
1.10 (0.76,1.59)
|
1.50 (1.18,1.91)
|
1.00 (0.76,1.33)
|
0.88 (0.71,1.09)
|
1.04 (0.81,1.34)
|
1.24 (0.92,1.65)
|
0.86 (0.69,1.06)
|
FI
|
|
1.40 (1.01,1.93)
|
|
0.93 (0.68,1.29)
|
1.27 (1.08,1.49)
|
0.96 (0.69,1.33)
|
0.84 (0.64,1.10)
|
1.00 (0.74,1.34)
|
1.18 (0.85,1.65)
|
0.82 (0.62,1.07)
|
0.96 (0.71,1.28)
|
AL
|
|
|
|
|
0.86 (0.66,1.13)
|
0.75 (0.62,0.92)
|
0.89 (0.71,1.13)
|
1.06 (0.80,1.40)
|
0.74 (0.60,0.90)
|
0.86 (0.68,1.08)
|
0.90 (0.68,1.19)
|
EX
|
|
0.67 (0.45,0.99)
|
0.91 (0.69,1.21)
|
1.08 (0.73,1.58)
|
0.94 (0.67,1.32)
|
1.12 (0.78,1.60)
|
1.32 (0.89,1.96)
|
0.92 (0.65,1.29)
|
1.07 (0.75,1.53)
|
1.12 (0.76,1.66)
|
1.25 (0.88,1.77)
|
SE
|
|
|
1.29 (0.97,1.72)
|
1.13 (0.90,1.41)
|
1.34 (1.04,1.73)
|
1.59 (1.18,2.13)
|
1.10 (0.88,1.38)
|
1.29 (1.00,1.65)
|
1.35 (1.00,1.81)
|
1.50 (1.19,1.89)
|
1.20 (0.84,1.72)
|
LI
|
1.14 (0.93,1.41)
|
0.89 (0.71,1.12)
|
0.78 (0.68,0.89)
|
0.92 (0.77,1.11)
|
1.09 (0.87,1.38)
|
0.76 (0.66,0.87)
|
0.89 (0.75,1.05)
|
0.93 (0.73,1.18)
|
1.03 (0.89,1.20)
|
0.83 (0.60,1.13)
|
0.69 (0.58,0.82)
|
PL
|
Note: EM: empagliflozin; CA: canagliflozin, DA: dapagliflozin, ER: ertugliflozin, SO: sotagliflozin, FI: finerenone, AL: albiglutide, EX: exenatide,
SE: semaglutide, LI: liraglutide, PL: placebo
|
Table 5b
Pairwise league table of renal outcomes.
Table 5b – Comparisons for renal outcome of the 8 drugs. RR with 95%CI
|
empagliflozin
|
1.07 (0.83,1.37)
|
0.91 (0.69,1.20)
|
1.26 (0.84,1.87)
|
1.31 (1.08,1.59)
|
1.38 (1.09,1.75)
|
1.50 (1.02,2.19)
|
1.34 (1.02,1.76)
|
1.53 (1.28,1.83)
|
0.94 (0.73,1.20)
|
canagliflozin
|
0.85 (0.65,1.13)
|
1.18 (0.79,1.76)
|
1.23 (1.01,1.50)
|
1.30 (1.02,1.65)
|
1.40 (0.96,2.06)
|
1.25 (0.95,1.65)
|
1.44 (1.20,1.72)
|
1.10 (0.83,1.45)
|
1.17 (0.89,1.55)
|
dapagliflozin
|
1.38 (0.91,2.09)
|
1.44 (1.14,1.81)
|
1.52 (1.16,1.98)
|
1.64 (1.10,2.45)
|
1.47 (1.09,1.98)
|
1.68 (1.36,2.08)
|
0.80 (0.53,1.19)
|
0.85 (0.57,1.27)
|
0.72 (0.48,1.10)
|
sotagliflozin
|
1.04 (0.72,1.51)
|
1.10 (0.74,1.63)
|
1.19 (0.73,1.95)
|
1.06 (0.70,1.61)
|
1.22 (0.85,1.74)
|
0.76 (0.63,0.93)
|
0.81 (0.67,0.99)
|
0.70 (0.55,0.87)
|
0.96 (0.66,1.39)
|
finerenone
|
1.06 (0.88,1.26)
|
1.14 (0.81,1.62)
|
1.02 (0.82,1.28)
|
1.17 (1.08,1.27)
|
0.72 (0.57,0.92)
|
0.77 (0.61,0.98)
|
0.66 (0.51,0.86)
|
0.91 (0.61,1.35)
|
0.95 (0.79,1.14)
|
dulaglutide
|
1.08 (0.75,1.57)
|
0.97 (0.74,1.26)
|
1.11 (0.94,1.30)
|
0.67 (0.46,0.98)
|
0.71 (0.49,1.04)
|
0.61 (0.41,0.91)
|
0.84 (0.51,1.37)
|
0.87 (0.62,1.24)
|
0.92 (0.64,1.34)
|
exenatide
|
0.89 (0.60,1.33)
|
1.02 (0.73,1.43)
|
0.75 (0.57,0.98)
|
0.80 (0.61,1.05)
|
0.68 (0.51,0.92)
|
0.94 (0.62,1.42)
|
0.98 (0.78,1.23)
|
1.03 (0.79,1.34)
|
1.12 (0.75,1.66)
|
liraglutide
|
1.14 (0.93,1.41)
|
0.65 (0.55,0.78)
|
0.70 (0.58,0.83)
|
0.59 (0.48,0.74)
|
0.82 (0.57,1.18)
|
0.86 (0.79,0.93)
|
0.90 (0.77,1.06)
|
0.98 (0.70,1.37)
|
0.87 (0.71,1.08)
|
placebo
|
Table 5c
Pairwise league table of ACD and CVD.
Table 5c – Comparisons for ACD (bottom left) and CVD (upper right) of the 8 drugs. RR with 95%CI
|
empagliflozin
|
1.11 (0.74,1.65)
|
1.09 (0.70,1.70)
|
1.12 (0.71,1.77)
|
1.14 (0.76,1.72)
|
1.13 (0.77,1.65)
|
1.40 (0.92,2.14)
|
0.88 (0.56,1.37)
|
1.28 (0.90,1.81)
|
1.02 (0.67,1.56)
|
canagliflozin
|
0.99 (0.71,1.38)
|
1.02 (0.72,1.44)
|
1.03 (0.77,1.37)
|
1.02 (0.80,1.30)
|
1.27 (0.94,1.72)
|
0.79 (0.57,1.11)
|
1.16 (0.95,1.40)
|
1.06 (0.69,1.62)
|
1.04 (0.78,1.37)
|
dapagliflozin
|
1.03 (0.69,1.54)
|
1.04 (0.74,1.48)
|
1.03 (0.76,1.41)
|
1.28 (0.89,1.84)
|
0.80 (0.54,1.18)
|
1.17 (0.89,1.54)
|
0.73 (0.41,1.30)
|
0.72 (0.45,1.16)
|
0.70 (0.43,1.12)
|
ertugliflozin
|
1.01 (0.71,1.46)
|
1.00 (0.72,1.39)
|
1.25 (0.86,1.81)
|
0.78 (0.52,1.16)
|
1.14 (0.85,1.52)
|
0.86 (0.57,1.31)
|
0.85 (0.65,1.10)
|
0.82 (0.63,1.06)
|
1.18 (0.74,1.87)
|
sotagliflozin
|
0.99 (0.76,1.28)
|
1.23 (0.90,1.69)
|
0.77 (0.54,1.09)
|
1.12 (0.91,1.39)
|
0.95 (0.64,1.41)
|
0.93 (0.74,1.16)
|
0.90 (0.72,1.13)
|
1.29 (0.83,2.02)
|
1.10 (0.90,1.35)
|
finerenone
|
1.24 (0.94,1.64)
|
0.78 (0.57,1.06)
|
1.13 (0.98,1.31)
|
0.84 (0.55,1.28)
|
0.82 (0.63,1.08)
|
0.79 (0.61,1.04)
|
1.14 (0.71,1.83)
|
0.97 (0.76,1.25)
|
0.88 (0.71,1.10)
|
exenatide
|
0.62 (0.43,0.90)
|
0.91 (0.72,1.15)
|
1.13 (0.73,1.73)
|
1.11 (0.83,1.47)
|
1.07 (0.80,1.42)
|
1.53 (0.95,2.48)
|
1.31 (1.00,1.70)
|
1.19 (0.94,1.50)
|
1.34 (1.02,1.77)
|
liraglutide
|
1.46 (1.11,1.92)
|
0.85 (0.58,1.24)
|
0.83 (0.69,1.02)
|
0.81 (0.66,0.98)
|
1.16 (0.75,1.79)
|
0.99 (0.83,1.17)
|
0.90 (0.80,1.00)
|
1.01 (0.84,1.22)
|
0.76 (0.62,0.93)
|
placebo
|
3.6. Conclusions from Minimally Contextualized Framework
The Minimally Contextualized Framework was developed to classify interventions in groups from the most to the least effective or harmful. The placebo was most closely connected to the other interventions and selected as the reference group, with an ineffective value, i.e. a relative effect value of 1, as the decision threshold. Based on the cardiovascular and renal outcomes,we used the 95% CI of the estimate of effect comparing each of the interventions against the placebo. If this interval crosses the decision threshold, then the intervention can remain in the same group as the placebo. On the other hand, if the interval did not cross the decision threshold, then depending on which side of the threshold the interval lies on, the intervention could be classified as more effective or less effective than the placebo. Based on comparisons between pairs of interventions, if any intervention proves to be more effective than another category 1 intervention, then that corresponding intervention can be moved to a higher rated group (category 2) [63]. After evaluating the certainty of the evidence fromall11 interventions, the interventions were classified again into two broad categories: high certainty (moderate to high certainty evidence) and low certainty (low to very low certainty evidence). After checking consistency with pairwise comparisons and rankings, the intervention at the highest classification level could be considered as the most effective choice currently available, while low certainty as might be among the most effective. As was presented in Table 6.
Table 6
Final classification of 11 interventions, based on NMA of interventions for patients with T2DM and CKD
|
Certainty of the evidence
|
Category
|
Intervention
|
Intervention vs placebo
RR (95% CI)
|
Surface under the cumulative ranking curve
|
MACE
|
|
|
|
|
High certainty (moderate to high certainty evidence)
|
Category 1: among the most effective
|
liraglutide
|
0.69 (0.58,0.82)
|
0.940
|
|
|
sotagliflozin
|
0.76 (0.66,0.87)
|
0.817
|
|
|
canagliflozin
|
0.78 (0.68,0.89)
|
0.772
|
|
Category 0: among the least effective
|
semaglutide
|
0.83 (0.60,1.13)
|
0.624
|
|
|
finerenone
|
0.89 (0.75,1.05)
|
0.508
|
|
|
empagliflozin
|
0.58 (0.25,1.36)
|
0.497
|
|
|
dapagliflozin
|
0.92 (0.77,1.11)
|
0.418
|
|
|
albiglutide
|
0.93 (0.73,1.18)
|
0.411
|
|
|
exenatide
|
1.03 (0.89,1.20)
|
0.173
|
|
|
ertugliflozin
|
1.09 (0.87,1.38)
|
0.113
|
Low certainty (low to very low certainty evidence)
|
Category 1: might be among the most effective
|
—
|
—
|
—
|
|
Category 0: might be among the least effective
|
—
|
—
|
—
|
Renal outcome
|
High certainty (moderate to high certainty evidence)
|
Category 2: among the most effective
|
dapagliflozin
|
0.59 (0.48,0.74)
|
0.941
|
|
|
empagliflozin
|
0.65 (0.55,0.78)
|
0.847
|
|
|
canagliflozin
|
0.70 (0.58,0.83)
|
0.765
|
|
Category 1: inferior to the most effective, or superior to the least effective
|
dapagliflozin
|
0.59 (0.48,0.74)
|
0.941
|
|
|
empagliflozin
|
0.65 (0.55,0.78)
|
0.847
|
|
|
canagliflozin
|
0.70 (0.58,0.83)
|
0.765
|
|
|
finerenone
|
0.86 (0.79,0.93)
|
0.437
|
|
Category 0: among the least effective
|
sotagliflozin
|
0.82 (0.57,1.18)
|
0.489
|
|
|
liraglutide
|
0.87 (0.71,1.08)
|
0.386
|
|
|
dulaglutide
|
0.90 (0.77,1.16)
|
0.321
|
|
|
exenatide
|
0.98 (0.70,1.37)
|
0.216
|
Low certainty (low to very low certainty evidence)
|
Category 1: might be among the most effective
|
—
|
—
|
—
|
|
Category 0: might be among the least effective
|
—
|
—
|
—
|
HHF
|
High certainty (moderate to high certainty evidence)
|
Category 2: among the most effective
|
ertugliflozin
|
0.51 (0.34,0.76)
|
0.863
|
|
Category 1: inferior to the most effective, or superior to the least effective
|
empagliflozin
|
0.59 (0.40,0.88)
|
0.702
|
|
|
ertugliflozin
|
0.51 (0.34,0.76)
|
0.863
|
|
|
dapagliflozin
|
0.56 (0.41,0.76)
|
0.785
|
|
|
canagliflozin
|
0.60 (0.48,0.75)
|
0.703
|
|
|
sotagliflozin
|
0.67 (0.52,0.85)
|
0.557
|
|
|
liraglutide
|
0.73 (0.56,0.97)
|
0.427
|
|
|
finerenone
|
0.79 (0.67,0.92)
|
0.327
|
|
Category 0: among the least effective
|
exenatide
|
1.10 (0.83,1.46)
|
0.040
|
Low certainty (low to very low certainty evidence)
|
Category 1: might be among the most effective
|
—
|
—
|
—
|
|
Category 0: might be among the least effective
|
—
|
—
|
—
|
ACD
|
|
|
|
|
High certainty (moderate to high certainty evidence)
|
Category 1: among the most effective
|
liraglutide
|
0.76 (0.62,0.93)
|
0.872
|
|
|
dapagliflozin
|
0.81 (0.66,0.98)
|
0.769
|
|
Category 0: among the least effective
|
canagliflozin
|
0.83 (0.69,1.02)
|
0.700
|
|
|
empagliflozin
|
0.85 (0.58,1.42)
|
0.624
|
|
|
finerenone
|
0.90 (0.80,1.00)
|
0.553
|
|
|
sotagliflozin
|
0.99 (0.83,1.17)
|
0.315
|
|
|
exenatide
|
1.01 (0.84,1.22)
|
0.258
|
|
|
ertugliflozin
|
1.16 (0.75,1.79)
|
0.156
|
Low certainty (low to very low certainty evidence)
|
Category 1: might be among the most effective
|
—
|
—
|
—
|
|
Category 0: might be among the least effective
|
—
|
—
|
—
|
CVD
|
|
|
|
|
High certainty (moderate to high certainty evidence)
|
Category 1: among the most effective
|
liraglutide
|
0.69 (0.52,0.90)
|
0.907
|
|
Category 0: among the least effective
|
empagliflozin
|
0.78 (0.55,1.11)
|
0.704
|
|
|
dapagliflozin
|
0.85 (0.65,1.13)
|
0.565
|
|
|
canagliflozin
|
0.87 (0.71,1.05)
|
0.554
|
|
|
finerenone
|
0.88 (0.76,1.02)
|
0.510
|
|
|
ertugliflozin
|
0.88 (0.66,1.18)
|
0.507
|
|
|
sotagliflozin
|
0.89 (0.72,1.10)
|
0.484
|
|
|
exenatide
|
1.10 (0.87,1.39)
|
0.091
|
Low certainty (low to very low certainty evidence)
|
Category 1: might be among the most effective
|
—
|
—
|
—
|
|
Category 0: might be among the least effective
|
—
|
—
|
—
|
Sensitivity analyses
The results of sensitivity analyses are summarized in Table 7. We conducted a sensitivity analysis excluding “Cherney 2021”, as Cherney 2021 only included diabetics with severe CKD (eGFR: 15-30 ml/min/1.73 m2). In MACE, renal outcomes and ACD, the results of sensitivity analyses were comparable to non-exclusion of “Cherney 2021”. Compared to sotagliflozin, liraglutide (RR [95% CI]; 0.76 [0.58–0.99]) was associated with a decreased risk of ACD. Whereas the previous results showed liraglutide had a trend towards a reduction in CVD compared to sotagliflozin.
Table 7
The summary of sensitivity analyses
|
Outcomes
|
Finerenone
|
SGLT-2i
|
GLP-1 RA
|
Comparison
|
Risk ratio
|
95%Cl
|
I2(%)
|
P
|
I
|
C
|
I
|
C
|
I
|
C
|
MACE
|
2833
|
2841
|
12959
|
11595
|
4674
|
4678
|
SGLT-2i vs placebo
|
0.84
|
0.78-0.91
|
26.8
|
0.174
|
|
|
|
|
|
|
|
GLP-1 RA vs placebo
|
0.88
|
0.80-0.97
|
|
|
|
|
|
|
|
|
|
Finerenone vs placebo
|
0.89
|
0.75-1.05
|
|
|
|
|
|
|
|
|
|
GLP-1 RA vs SGLT-2i
|
1.04
|
0.92-1.18
|
|
|
|
|
|
|
|
|
|
Finerenone vs SGLT-2i
|
1.05
|
0.87-1.27
|
|
|
|
|
|
|
|
|
|
Finerenone vs GLP-1 RA
|
1.01
|
0.83-1.23
|
|
|
HHF
|
6519
|
6507
|
12960
|
11596
|
2673
|
2662
|
SGLT-2i vs placebo
|
0.60
|
0.53-0.68
|
49.9
|
0.030
|
|
|
|
|
|
|
|
GLP-1 RA vs placebo
|
0.90
|
0.73-1.09
|
|
|
|
|
|
|
|
|
|
Finerenone vs placebo
|
0.79
|
0.67-0.92
|
|
|
|
|
|
|
|
|
|
GLP-1 RA vs SGLT-2i
|
1.49
|
1.18-1.89
|
|
|
|
|
|
|
|
|
|
Finerenone vs SGLT-2i
|
1.31
|
1.07-1.61
|
|
|
|
|
|
|
|
|
|
Finerenone vs GLP-1 RA
|
0.88
|
0.68-1.14
|
|
|
Renal
outcome
|
7234
|
6600
|
11547
|
10888
|
3754
|
3780
|
SGLT-2i vs placebo
|
0.66
|
0.59-0.73
|
39.2
|
0.079
|
|
|
|
|
|
|
|
GLP-1 RA vs placebo
|
0.90
|
0.80-1.02
|
|
|
|
|
|
|
|
|
|
Finerenone vs placebo
|
0.86
|
0.79-0.93
|
|
|
|
|
|
|
|
|
|
GLP-1 RA vs SGLT-2i
|
1.37
|
1.17-1.61
|
|
|
|
|
|
|
|
|
|
Finerenone vs SGLT-2i
|
1.30
|
1.14-1.49
|
|
|
|
|
|
|
|
|
|
Finerenone vs GLP-1 RA
|
0.95
|
0.82-1.10
|
|
|
CVD
|
6519
|
6507
|
12960
|
11596
|
2673
|
2662
|
SGLT-2i vs placebo
|
0.87
|
0.78-0.97
|
0.0
|
0.658
|
|
|
|
|
|
|
|
GLP-1 RA vs placebo
|
0.90
|
0.75-1.08
|
|
|
|
|
|
|
|
|
|
Finerenone vs placebo
|
0.88
|
0.76-1.02
|
|
|
|
|
|
|
|
|
|
GLP-1 RA vs SGLT-2i
|
1.04
|
0.84-1.28
|
|
|
|
|
|
|
|
|
|
Finerenone vs SGLT-2i
|
1.02
|
0.85-1.22
|
|
|
|
|
|
|
|
|
|
Finerenone vs GLP-1 RA
|
0.98
|
0.78-1.23
|
|
|
ACD
|
6519
|
6507
|
11944
|
10805
|
2673
|
2662
|
SGLT-2i vs placebo
|
0.90
|
0.81-0.99
|
0.0
|
0.537
|
|
|
|
|
|
|
|
GLP-1 RA vs placebo
|
0.89
|
0.77-1.02
|
|
|
|
|
|
|
|
|
|
Finerenone vs placebo
|
0.90
|
0.80-1.00
|
|
|
|
|
|
|
|
|
|
GLP-1 RA vs SGLT-2i
|
0.99
|
0.83-1.17
|
|
|
|
|
|
|
|
|
|
Finerenone vs SGLT-2i
|
1.00
|
0.86-1.16
|
|
|
|
|
|
|
|
|
|
Finerenone vs GLP-1 RA
|
1.01
|
0.85-1.20
|
|
|
Note: I: intervention; C: control
|