Effect of sodium-glucose cotransporter-2 inhibitors on cardiac remodelling: a systematic review and meta-analysis

Abstract

Aims
To examine the effects of sodium-glucose cotransporter-2 inhibitors (SGLT2i) on cardiac remodelling in patients with type 2 diabetes mellitus (T2DM) and/or heart failure (HF), and to explore the subsets of patients who may have greater benefit from SGLT2i therapy.

Methods and results
Four electronic databases were searched for randomized controlled trials (RCTs) that evaluated the effects of SGLT2i on parameters reflecting cardiac remodelling in patients with T2DM and/or HF. Standardized mean differences (SMDs) or mean differences (MDs) were pooled. Subgroup analyses were performed according to the baseline HF and T2DM, HF type, SGLT2i agent, follow-up duration, and imaging modality. A total of 13 RCTs involving 1251 patients were analysed. Sodium-glucose cotransporter-2 inhibitors treatment significantly improved left ventricular (LV) ejection fraction [SMD, 0.35; 95% confidence interval (CI) (0.04, 0.65); P = 0.03], LV mass [SMD, −0.48; 95% CI (−0.79, −0.18); P = 0.002], LV mass index [SMD, −0.27; 95% CI (−0.49, −0.05); P = 0.02], LV end-systolic volume [SMD, −0.37; 95% CI (−0.71; −0.04); P = 0.03], LV end-systolic volume index [MD, −0.35 mL/m2; 95% CI (−0.64, −0.05); P = 0.02], and E-wave deceleration time [SMD, −0.37; 95% CI (−0.70, −0.05); P = 0.02] in the overall population. Subgroup analyses showed that the favourable effects of SGLT2i on LV remodelling were only significant in HF patients, especially HF with reduced ejection fraction (HFrEF), regardless of glycaemic status. Among the four included SGLT2i, empagliflozin was associated with a greater improvement of LV mass, LV mass index, LV end-systolic volume, LV end-systolic volume index, LV end-diastolic volume, and LV end-diastolic volume index (all P < 0.05).

Conclusions
Sodium-glucose cotransporter-2 inhibitors treatment significantly reversed cardiac remodelling, improving LV systolic and diastolic function, LV mass and volume, especially in patients with HFrEF and amongst those taking empagliflozin compared with other SGLT2i. Reversed remodelling may be a mechanism responsible for the favourable clinical effects of SGLT2i on HF.

Full text

.
.
.
.
Effect of sodium-glucose cotransporter-2
inhibitors on cardiac remodelling: a systematic
review and meta-analysis
Nan Zhang
1†
, Yueying Wang
1†
, Gary Tse
1,2,3
, Panagiotis Korantzopoulos
4
,
Konstantinos P. Letsas
5
, Qingpeng Zhang
6
, Guangping Li
1
, Gregory Y.H. Lip
7
, and
Tong Liu
1
*
‡
1
Tianjin Key Laboratory of Ionic-Molecular Function of Cardiovascular Disease, Department of Cardiology, Tianjin Institute of Cardiology, Second Hospital of Tianjin Medical
University, Tianjin 300211, China;
2
Kent and Medway Medical School, Canterbury, Kent CT2 7NT, UK;
3
Faculty of Health and Medical Sciences, University of Surrey, Guildford
GU2 7AL, UK;
4
First Department of Cardiology, University of Ioannina Medical School, Ioannina, Greece;
5
Arrhythmia Unit, Laboratory of Cardiac Pacing and Electrophysiology,
Onassis Cardiac Surgery Center, Athens, Greece;
6
School of Data Science, City University of Hong Kong, Hong Kong, China; and
7
Liverpool Centre for Cardiovascular Science,
University of Liverpool and Liverpool Heart & Chest Hospital, Liverpool, UK
Received 10 August 2021; accepted 4 October 2021
Aims To examine the effects of sodium-glucose cotransporter-2 inhibitors (SGLT2i) on cardiac remodelling in patients
with type 2 diabetes mellitus (T2DM) and/or heart failure (HF), and to explore the subsets of patients who may
have greater benefit from SGLT2i therapy.
........................................................................ ............. ............. ............. .................. ..................................................................
Methods
and results
Four electronic databases were searched for randomized controlled trials (RCTs) that evaluated the effects of
SGLT2i on parameters reflecting cardiac remodelling in patients with T2DM and/or HF. Standardized mean differ-
ences (SMDs) or mean differences (MDs) were pooled. Subgroup analyses were performed according to the base-
line HF and T2DM, HF type, SGLT2i agent, follow-up duration, and imaging modality. A total of 13 RCTs involving
1251 patients were analysed. Sodium-glucose cotransporter-2 inhibitors treatment significantly improved left ven-
tricular (LV) ejection fraction [SMD, 0.35; 95% confidence interval (CI) (0.04, 0.65); P= 0.03], LV mass [SMD,
0.48; 95% CI (0.79, 0.18); P= 0.002], LV mass index [SMD, 0.27; 95% CI (0.49, 0.05); P= 0.02], LV end-
systolic volume [SMD, 0.37; 95% CI (0.71; 0.04); P= 0.03], LV end-systolic volume index [MD, 0.35 mL/m
2
;
95% CI (0.64, 0.05); P= 0.02], and E-wave deceleration time [SMD, 0.37; 95% CI (0.70, 0.05); P= 0.02] in
the overall population. Subgroup analyses showed that the favourable effects of SGLT2i on LV remodelling were
only significant in HF patients, especially HF with reduced ejection fraction (HFrEF), regardless of glycaemic status.
Among the four included SGLT2i, empagliflozin was associated with a greater improvement of LV mass, LV mass
index, LV end-systolic volume, LV end-systolic volume index, LV end-diastolic volume, and LV end-diastolic volume
index (all P< 0.05).
........................................................................ ............. ............. ............. .................. ..................................................................
Conclusions Sodium-glucose cotransporter-2 inhibitors treatment significantly reversed cardiac remodelling, improving LV sys-
tolic and diastolic function, LV mass and volume, especially in patients with HFrEF and amongst those taking empa-
gliflozin compared with other SGLT2i. Reversed remodelling may be a mechanism responsible for the favourable
clinical effects of SGLT2i on HF.
䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏䊏
* Corresponding author. Tel: þ86-22-88328617, Email: liutong@tmu.edu.cn;liutongdoc@126.com
†
These authors are co-first authors and contributed equally to the study.
‡
This author is joint senior author.
Published on behalf of the European Society of Cardiology. All rights reserved. V
CThe Author(s) 2021. For permissions, please email: journals.permissions@oup.com.
European Journal of Preventive Cardiology FULL RESEARCH PAPER
https://doi.org/10.1093/eurjpc/zwab173
Downloaded from https://academic.oup.com/eurjpc/advance-article/doi/10.1093/eurjpc/zwab173/6430917 by guest on 19 November 2021

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
........................................................................ ............. ............. ............. .................. ..................................................................
........................................................................ ............. ............. ............. .................. ..........................................................
Keywords Sodium-glucose cotransporter-2 inhibitors •Cardiac remodelling •Heart failure •HFrEF •Type 2 diabetes
mellitus
Introduction
Sodium-glucose cotransporter-2 inhibitors (SGLT2i) represent
the latest class of anti-diabetic agents and have received significant
attention owing to their beneficial effects on cardiovascular
events, especially in the context of heart failure (HF). Recent clin-
ical trials have demonstrated that SGLT2i treatment was associ-
ated with a significant reduction of cardiovascular death and HF
hospitalization risk in patients with type 2 diabetes mellitus
(T2DM),
1–3
or those with HF with reduced ejection fraction
(HFrEF) with or without accompanying T2DM.
4,5
Whilst previous
systematic reviews and meta-analyses on SGLT2i have been per-
formed for clinical outcomes
6–10
or on biomarkers of inflamma-
tion and oxidative stress,
11
a thorough evaluation on its effects on
cardiac remodelling has not been performed.
Cardiac remodelling is a pivotal mechanism for HF development
and progression and is associated with poor clinical outcomes.
Remodelling is defined as changes in cardiac geometry and/or func-
tion which can be measured through changes of cardiac chamber
dimensions, volumes, mass, and functions, mostly by echocardiog-
raphy or cardiac magnetic resonance (CMR).
12–14
Several experi-
mental studies have demonstrated the beneficial effects of SGLT2i on
cardiac remodelling.
15–18
In the SUGAR-DM-HF trial conducted in
patients with HFrEF and T2DM or pre-diabetes, SGLT2i was also
associated with a significant reduction in left ventricular (LV) volumes
compared with placebo.
19
In contrast, the REFORM trial failed to
demonstrate any effect of SGLT2i on cardiac remodelling in patients
with HF and T2DM.
20
Given the discrepant findings from the existing
studies, we conducted a systematic review and meta-analysis of
randomized controlled trials (RCTs) to evaluate the effects of
SGLT2i treatment on cardiac remodelling in patients with T2DM
and/or HF. Furthermore, comprehensive subgroup analyses were
performed to explore the subsets of patients who may benefit more
from SGLT2i therapy.
Methods
Eligibility criteria
This meta-analysis was performed in accordance with the Preferred
Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)
guidelines. An RCT was eligible if the following criteria were fulfilled: (i)
patients aged 18 years or older with a diagnosis of T2DM and/or HF; (ii)
comparison between SGLT2i and placebo or active control; (iii) reported
at least one outcome variable assessed by CMR or echocardiography.
The outcomes of this meta-analysis were the change of remodelling
parameters of LV and right ventricular (RV), including systolic and diastol-
ic function, mass, and volume.
Graphical Abstract
2N. Zhang et al.
Downloaded from https://academic.oup.com/eurjpc/advance-article/doi/10.1093/eurjpc/zwab173/6430917 by guest on 19 November 2021

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Search strategy
A systematic literature search of relevant RCTs was conducted in
PubMed, Embase, Cochrane Library, and ClinicalTrials.gov from inception
through 28 February 2021, without any restriction. We screened pub-
lished and unpublished RCTs using a strategy comprising both MeSH
terms and free text. The detailed search algorithm is presented in
Supplementary material online, S1. References listed in the identified
studies were scrutinized for relevant studies. Trial eligibility was con-
firmed by two independent reviewers (N.Z. and Y.W.) and discrepancies
were resolved by a third author (T.L.).
Data extraction and quality assessment
Two reviewers (N.Z. and Y.W.) independently performed data ex-
traction and quality assessment. Pre-specified forms were applied to
collect data including the following items: (i) first author, publication
year, number of patients, study design, and follow-up duration; (ii)
characteristics of patients; (iii) details of SGLT2i and control, such as
type and dosage; and (iv) outcome-related parameters. The methodo-
logical quality of the included studies was evaluated using the
Cochrane Risk of Bias Tool.
Statistical analysis
Outcome variables reported by at least two studies were eligible to
be analysed. We used mean (SD) change from baseline to calculate
the pooled effects. Other forms of results, such as confidence inter-
vals (CIs), were transformed to SDs according to Cochrane
Handbook. If only median values with interquartile range were
reported, means and SDs were estimated using the Box-Cox
method.
21
When change-from-baseline mean (SD) was missing, but
baseline and final measurements were reported separately, we calcu-
lated the correlation coefficient according to other included studies
which provided mean (SD) for baseline, final, and change values. A
conservative estimate (minimum correlation) was used to impute the
missing SD of mean changes.
22
Correlations varied from 0.69 to 0.88
in this meta-analysis. One included cross-over study (the DapKid)
only reported the final measurements, considering the identical base-
line characteristics between the intervention and control arms, we
combined the final values from this study with other change-from-
baseline values directly.
23
For outcomes where change-from-baseline
values and correlation coefficient were both unavailable, we com-
bined final values directly. Sensitivity analyses were performed by
omitting studies with imputed SDs and final values to assess the ro-
bustness of the results.
Due to the different imaging methods used in the included studies,
pooled effects were summarized as standardized mean differences
(SMDs) with corresponding 95% CIs according to the inverse variance
method. Mean differences (MDs) were used when combined final values
with change-from-baseline values. A random-effects model was applied.
Heterogeneity was assessed by using the Cochrane Q statistic and I
2
stat-
istic. For the Q test, P-value <0.1 was considered statistically significant. I
2
>70% was considered as high heterogeneity and I
2
>50% as moderate.
We used intention-to-treat data wherever possible.
Predefined subgroup analyses were conducted by different clinical
conditions (with or without T2DM or HF), HF types [HFrEF and heart
failure with preserved ejection fraction (HFpEF)], SGLT2i types, control
types, follow-up durations, and imaging methods. To evaluate the impact
of each study on the overall effect size, a one-study removed sensitivity
analysis was performed. Publication bias was visually assessed by funnel
plot, but no further testing was performed given the low number of
included studies. A P-value <0.05 was considered significant. Statistical
analyses were conducted with RevMan version 5.3 and R version 4.0.4.
Results
Study selection and study characteristics
Initially, 315 records were identified, and subsequently, 13 RCTs
were included in this meta-analysis.
19,20,23–33
A detailed flowchart for
study selection is presented in Figure 1. The characteristics of the
included studies are summarized in Table 1. In total, 1251 patients
were included, of whom 638 were treated with SGLT2i and 649
served as controls (36 patients with cross-over). The mean age of
participants was 56.0 to 73.1 years, and the proportion of males
ranged from 53.0% to 93.0%, with a mean follow-up duration ranging
from 6 weeks to 1 year. The numbers of studies and patients available
in each variable are presented in Table 2.
Nine RCTs
19,20,23–28,30
were placebo-controlled, others
29,31–33
were active-controlled. Four RCTs
19,24–26
evaluated the effect of
empagliflozin, six of dapagliflozin,
20,23,27–30
two of canagliflozin,
31,32
and one of luseogliflozin.
33
Cardiac magnetic resonance and echocar-
diographywereusedinfour
19,20,25,30
and seven RCTs,
23,26,28,29,31–33
respectively, whereas outcome variables of the remaining two
RCTs
24,27
were from both the CMR main-study and the pre-specified
echocardiographic sub-study. Except for the EMPA-TROPISM
(ATRU-4) study
25
(all patients without T2DM) and the Empire HF
study
26
(12.6% with T2DM), all studies limited their participants to
the T2DM population. Seven RCTs included patients with established
HF, four of which with HFrEF,
19,25,26,32
one RCT with HFpEF,
33
two
RCTs
20,31
included both types but only one provided the separate
results. The study population in six RCTs
23,24,27–30
were without or
not limited to HF. Results of the overall meta-analyses excluding and
including the imputed studies are presented in Table 2.
Risk of bias assessment
Details about the risk of bias assessment are shown in
Supplementary material online, S2. The overall risk of bias was found
to be low in most RCTs.
Meta-analyses of left ventricular systolic
function and structure
Left ventricular ejection fraction
In the overall analysis without the imputed studies, SGLT2i treatment
significantly improved the LV ejection fraction (LVEF) compared with
placebo and active control [SMD, 0.35; 95% CI (0.04, 0.65); P=0.03]
(Figure 2A) or only compared with placebo [SMD, 0.39; 95% CI (0.01,
0.77); P= 0.04] in patients with T2DM and/or HF (Supplementary
material online, Figure S3A). After including the imputed studies, the
favourable effect of SGLT2i was still significant in the overall popula-
tion [MD, 1.49%; 95% CI (0.15, 2.84); P= 0.03] (Supplementary ma-
terial online, Figure S4A).
Left ventricular mass and left ventricular mass index
The aggregated data of four non-imputed RCTs showed that SGLT2i
treatment significantly reduced the LV mass (LVM) compared with
placebo [SMD, 0.48; 95% CI (0.79, 0.18); P= 0.002] in patients
with T2DM and/or HF (Figure 2B). When indexed to body surface
area, the favourable effect of SGLT2i was still significant. In the overall
analysis without imputed studies, SGLT2i treatment was associated
with a significant reduction of left ventricular mass index (LVMi) vs.
SGLT2 inhibitors and cardiac remodelling 3
Downloaded from https://academic.oup.com/eurjpc/advance-article/doi/10.1093/eurjpc/zwab173/6430917 by guest on 19 November 2021

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
placebo [SMD, 0.27; 95% CI (0.49, 0.05); P= 0.02] (Figure 2C).
When including the imputed studies, SGLT2i treatment was associ-
ated with improved LVMi but did not reach the statistical significance
[MD, 1.78 g/m
2
; 95% CI (3.55, 0.02); P=0.05] (Supplementary
material online, Figure S13A).
Left ventricular end-systolic volume and left ventricular
end-systolic volume index
After pooling seven non-imputed RCTs, a significant association be-
tween SGLT2i therapy and LV end-systolic volume (LVESV) reduc-
tion was found in the overall population compared with placebo
[SMD, 0.37; 95% CI (0.71; 0.04); P=0.03] (Figure 3A). Left ven-
tricular end-systolic volume index (LVESVi) was significantly reduced
after treatment with SGLT2i [SMD, 0.35; 95% CI (0.64, 0.05);
P=0.02] (Supplementary material online, Figure S25A) or exhibited a
borderline significance [SMD, 0.38; 95% CI (0.76, 0.01); P=0.05]
(Figure 3B), when including or excluding the imputed studies,
respectively.
Left ventricular end-diastolic volume and left ventricular
end-diastolic volume index
Sodium-glucose cotransporter-2 inhibitors treatment was not associ-
ated with a significant reduction of LV end-diastolic volume (LVEDV)
[SMD, 0.28; 95% CI (0.60, 0.05); P= 0.10] or LVEDVi [SMD,
0.35; 95% CI (0.80, 0.11); P= 0.13] compared with placebo in
patients with T2DM and/or HF (Figure 3C and D).
Left ventricular global longitudinal strain
In the pooled analysis without imputed studies, there was no signifi-
cant difference in mean change of LV global longitudinal strain
(LVGLS) between SGLT2i and placebo [SMD, 0.09; 95% CI (0.50,
0.33), P=0.68] (Figure 2D), which was consistent after including the
imputed studies [MD, 0.11%; 95% CI (0.62, 0.41); P=0.68]
(Supplementary material online, Figure S36).
Meta-analyses of left ventricular diastolic
function
E/e0and left atrial volume index
The aggregated results of five non-imputed studies [SMD, 0.39;
95% CI (0.82, 0.05); P=0.08](Figure 4A) and eight studies (including
both imputed and non-imputed data) [MD, 0.71; 95% CI (1.44,
0.03); P= 0.06] (Supplementary material online, Figure S40A) both
indicated a signal of improved E/e0in patients receiving SGLT2i, but
the differences between groups did not meet the conventional level
of statistical significance. After removing the MUSCAT-HF
33
from
Japan that used luseogliflozin, E/e0was significantly reduced after
SGLT2i treatment [MD, 0.84; 95% CI (1.60, 0.09); P=0.03]
(Supplementary material online, Figure S41). As for left atrial volume
index (LAVi), the aggregated results indicated that SGLT2i therapy
did not significantly improve LAVi compared with placebo in the
overall population [SMD, 0.14; 95% CI (0.32, 0.04); P=0.12]
(Figure 4B).
Septal e0and lateral e0
There was no significant difference in change of septal e0[SMD,
0.22; 95% CI (0.47, 0.03); P= 0.09] (Figure 4C)andlaterale
0
Figure 1 Flow diagram of the study selection process.
4N. Zhang et al.
Downloaded from https://academic.oup.com/eurjpc/advance-article/doi/10.1093/eurjpc/zwab173/6430917 by guest on 19 November 2021

..................................................................................................................................................................................................................................................................................................
Table 1 Characteristics of the studies included in meta-analysis
Study, year Study design Condition Number Age Males Country Type of
control
Type of
SGLT2i
Follow-up
duration
Imaging
method
Outcome variables
(unique identifier) (SGLT2i/
control)
(years) (%)
EMPA-HEART
CardioLink-
6,
24
2019
RCT T2DM and
CAD
97 (49/48) 62.8 ± 9.0 93.0 Canada Placebo Empagliflozin 6 months Echo, CMR LVEF, LVM, LVMi,
LVEDV, LVEDVi,
LVESV, LVESVi, E/
e0, LAVi
Original
a
(NCT02998970) (6% with HF) (10 mg/day)
EMPA-TROPISM
(ATRU-4),
25
2020
RCT HFrEF without
T2DM
84 (42/42) 62 ± 12.1 64.0 USA Placebo Empagliflozin 6 months CMR LVEF, LVM, LVMi,
LVEDV, LVEDVi,
LVESV, LVESVi
Original
(NCT03485222) (10 mg/day)
SUGAR-DM-
HF,
19
2020
RCT HFrEF and
T2DM or
pre-diabetes
105 (52/53) 68.7 ± 11.1 73.3 UK Placebo Empagliflozin 36 weeks CMR LVEF, LVM, LVMi,
LVEDV, LVEDVi,
LVESV, LVESVi,
LAVi, LVGLS
Original
(NCT03485092) (10 mg/day)
Empire HF,
26
2021
RCT HFrEF (12.6%
with T2DM)
190 (95/95) 64 ± 11 85.3 Denmark Placebo Empagliflozin 12 weeks Echo LVEF, LVMi, LVEDV,
LVEDVi, LVESV,
LVESVi, LAVi,
LVGLS
Original
(NCT03198585) (10 mg/day)
DAPA-LVH,
27
2020
RCT T2DM without
clinical HF
66 (32/34) 65.5 ± 6.87 57.6 UK Placebo Dapagliflozin 1 year Echo, CMR LVEF, LVM, LVMi,
LVEDV, LVESV,
DT, lateral e0, septal
e0, E/e0, LVGLS
Original
(NCT02956811) (10 mg/day)
IDDIA,
28
2020 RCT T2DM 60 (30/30) NA NA Korea Placebo Dapagliflozin 24 weeks Echo LVMi, LAVi, E/e0Original
(NCT02751398) (10 mg/day)
REFORM,
20
2020 RCT T2DM and HF 56 (28/28) 67.1 ± 6.9 66.1 UK Placebo Dapagliflozin 1 year CMR LVEF, LVMi, LVEDV,
LVEDVi, LVESV,
LVESVi, LAVi
Original
(NCT02397421) (10 mg/day)
IDOL-EPOC,
29
2020
RCT T2DM 74 (36/38) 67.7 ± 8.5 89.2 Japan Conventional
therapies
Dapagliflozin 6 months Echo E/e0Original
(UMIN000023834) (5 mg/day)
DapKid,
23
2020
a
RCT T2DM and
albuminuria
36 (36/36)
b
64 ± 8 89.0 Denmark Placebo Dapagliflozin 12 weeks Echo LVEF, LVMi, E/e0, E/A,
LVGLS, TAPSE
Imputed
(NCT02914691) (10 mg/day)
DAPACARD,
30
2021
RCT T2DM without
HF
49 (25/24) 64.4 ± 7.2 53.0 Sweden
and
Placebo Dapagliflozin 6 weeks CMR LVEF, LVMi, LVEDVi,
LVESVi, LVGLS,
DT, E/A
Imputed
(NCT03387683) Finland (10 mg/day)
CANDLE,
31
2020 RCT T2DM and HF 233 (113/120) 68.6 ± 10.1 74.6 Japan Limepiride Canagliflozin 24 weeks Echo LVEF, E/e0, septal e0,
lateral e0
Original
(UMIN000017669) (100 mg/day)
CANA-HF,
32
2020
RCT T2DM and
HFrEF
36 (17/19) 56.0 ± 7.8 77.8 USA Sitagliptin Canagliflozin 12 weeks Echo LVEF, LVEDVi,
LVESVi, E/e0, DT,
TAPSE
Imputed
(NCT02920918) (100 mg/day)
MUSCAT-HF,
33
2020
RCT T2DM and
HFpEF
165 (83/82) 73.1 ± 7.8 62.4 Japan Voglibose Luseogliflozin 12 weeks Echo LVEF, LVMi, E/e0, E/A,
LAVi
Imputed
(UMIN000018395) (2.5 mg/day)
CAD, coronary artery disease; CMR, cardiac magnetic resonance; DT, E-wave deceleration time; E/A, ratio of early to atrial mitral inflow velocity; Echo, echocardiography; E/e0, ratio of early mitral inflow velocity (E) to early mitral annular
velocity (e0); HF, heart failure; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; lateral e0, early lateral annular tissue Doppler velocity; LAVi, left atrial volume index; LVEDV, left ven-
tricular end-diastolic volume; LVEDVi, left ventricular end-diastolic volume index; LVEF, left ventricular ejection fraction; LVESV, left ventricular end-systolic volume; LVESVi, left ventricular end-systolic volume index; LVGLS, left ventricular
global longitudinal strain; LVM, left ventricular mass; LVMi, left ventricular mass index; NA, not available; RCT, randomized controlled trial; septal e0, early septal annular tissue Doppler velocity; SGLT2i, sodium-glucose cotransporter-2 inhib-
itors; T2DM, type 2 diabetes mellitus; TAPSE, tricuspid annular plane systolic excursion.
a
Original means outcome variables are extracted from the original study but not imputed.
b
Participants in the DapKid study crossed over to the opposite treatment after first 12 weeks.
SGLT2 inhibitors and cardiac remodelling 5
Downloaded from https://academic.oup.com/eurjpc/advance-article/doi/10.1093/eurjpc/zwab173/6430917 by guest on 19 November 2021

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
[SMD, 0.09; 95% CI (0.41, 0.22); P= 0.57] (Figure 4D)between
SGLT2i and controls in patients with T2DM and/or HF.
E/A and deceleration time
The pooled results of three studies showed no significant association
between SGLT2i therapy and E/A (MD, 0.01; 95% CI [0.07, 0.09];
P=0.78) (Figure 4E). As for E-wave deceleration time (DT), the
pooled result indicated that SGLT2i treatment was associated with a
significant reduction of DT in the overall population [SMD, 0.37;
95% CI (0.70, 0.05); P= 0.02] (Figure 4F). Although E/A and DT
are heart rate-dependent and were not recommended as diagnostic
criteria of HFpEF in the recent European HF guidelines,
34,35
both
parameters are important measurements of mitral inflow and stand-
ard measurements of LV diastolic function. Therefore, significant im-
provement of DT after SGLT2i treatment might also suggest a
favourable effect of SGLT2i on LV diastolic function.
Meta-analysis of right ventricular
function
Tricuspid annular plane systolic excursion
After combining the final values (as opposed to the change-from-
baseline values) of tricuspid annular plane systolic excursion (TAPSE)
reported by two studies, the results indicate that SGLT2i treatment
was not associated with a significant change of TAPSE [MD, 0.11 cm;
95% CI (0.06, 0.29); P=0.21] (Supplementary material online,
Figure S50).
Results of subgroup analyses
To identify the subsets of patients who may benefit more from
SGLT2i therapy, subgroup analyses were conducted according to
baseline HF. Treatment with SGLT2i significantly improved the LVEF
[MD, 2.08%; 95% CI (0.06, 4.11); P= 0.04], LVESVi [SMD, 0.41; 95%
CI (0.82, 0.01); P= 0.04] and LAVi [MD, 1.98 mL/m
2
;95%CI
(3.86, 0.10); P=0.04], but not LVM [SMD, 0.53; 95% CI (1.21,
0.16); P= 0.13], LVMi [MD, 4.19 g/m
2
; 95% CI (9.07, 0.69);
P= 0.09], LVESV [SMD, 0.45; 95% CI (0.95, 0.05); P= 0.08],
LVEDV [SMD, 0.39; 95% CI (0.83, 0.04); P= 0.08], LVEDVi [SMD,
0.33; 95% CI (0.81, 0.16); P= 0.19], LVGLS [MD, 0.21%; 95% CI
(0.25, 0.67); P= 0.37], or E/e0[MD, 0.00; 95% CI (1.01, 1.01);
P= 1.00] in patients with HF. Among the patients without HF or not
limited to HF, SGLT2i therapy failed to demonstrate a significant ef-
fect on these variables (Supplementary material online, S3).
Further subgroup analyses according to HF phenotypes showed
that SGLT2i treatment significantly improved the LVEF in HFrEF
patients [SMD, 0.52; 95% CI (0.04, 1.01); P=0.04], but not in HFpEF
patients [SMD, 0.01; 95% CI (0.21, 0.24); P= 0.90]. Left ventricular
mass index was significantly reduced after SGLT2i therapy in patients
with HFrEF [SMD, 0.42; 95% CI (0.75, 0.09); P= 0.01], but not
in patients with HFpEF [SMD, 0.21; 95% CI (0.51, 0.10); P= 0.19].
..................................................................................... .......................................................................................
....................................................................................................................................................................................................................
Table 2 Results of overall meta-analyses comparing between sodium-glucose cotransporter-2 inhibitors and control
Measurements Overall analyses without the imputed studies Overall analyses including the imputed studies
Studies (n) SMD (95%CI) P-value Studies (n) MD or SMD (95%CI)
a
P-value
LVEF 7 (759) 0.35 (0.04, 0.65) 0.03 11 (1079) 1.49% (0.15, 2.84) 0.03
LVM
b
4 (328) 0.48 (0.79, 0.18) 0.002 —
LVMi 7 (621) 0.27 (0.49, 0.05) 0.02 10 (905) 1.78 g/m
2
(3.55, 0.02) 0.05
LVESV
b
6 (563) 0.37 (0.71, 0.04) 0.03 —
LVESVi 5 (497) 0.38 (0.76, 0.01) 0.05 7 (582) 0.35 (0.64, 0.05) 0.02
LVEDV
b
6 (563) 0.28 (0.60, 0.05) 0.10 —
LVEDVi 5 (497) 0.35 (0.80, 0.11) 0.13 7 (582) 0.21 (0.59, 0.16) 0.27
LVGLS 3 (318) 0.09 (0.50, 0.33) 0.68 5 (435) 0.11% (0.62, 0.41) 0.68
E/e05 (450) 0.39 (0.82, 0.05) 0.08 8 (721) 0.71 (1.44, 0.03) 0.06
E/A
c
— 3 (284) 0.01 (0.07, 0.09) 0.78
DT
c
— 3 (147) 0.37 (0.70, 0.05) 0.02
LAVi 5 (482) 0.14 (0.32, 0.04) 0.12 6 (647) 1.04 mL/m
2
(2.35, 0.27) 0.12
Septal e0
b
2 (247) 0.22 (0.47, 0.03) 0.09 —
Lateral e0
b
2 (223) 0.09 (0.41, 0.22) 0.57 —
TAPSE
c
— 2 (106) 0.11 cm (0.06, 0.29) 0.21
P-values in bold indicate <0.05.
CAD, coronary artery disease; CI, confidence interval; CMR, cardiac magnetic resonance; DT, E-wave deceleration time; E/A, ratio of early to atrial mitral inflow velocity; Echo,
echocardiography; E/e0, ratio of early mitral inflow velocity (E) to early mitral annular velocity (e0); HF, heart failure; HFpEF, heart failure with preserved ejection fraction;
HFrEF, heart failure with reduced ejection fraction; lateral e0, early lateral annular tissue Doppler velocity; LAVi, left atrial volume index; LVEDV, left ventricular end-diastolic
volume; LVEDVi, left ventricular end-diastolic volume index; LVEF, left ventricular ejection fraction; LVESV, left ventricular end-systolic volume; LVESVi, left ventricular end-sys-
tolic volume index; LVGLS, left ventricular global longitudinal strain; LVM, left ventricular mass; LVMi, left ventricular mass index; MD, mean difference; NA, not available; RCT,
randomized controlled trial; septal e0, early septal annular tissue Doppler velocity; SGLT2i, sodium-glucose cotransporter-2 inhibitors; SMD, standardized mean difference;
T2DM, type 2 diabetes mellitus; TAPSE, tricuspid annular plane systolic excursion.
a
MD was used when final values were combined with change-from-baseline values.
b
Outcomes variables without the imputed study.
c
Outcomes variables only including the imputed studies (DT including one non-imputed study).
6N. Zhang et al.
Downloaded from https://academic.oup.com/eurjpc/advance-article/doi/10.1093/eurjpc/zwab173/6430917 by guest on 19 November 2021

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Sodium-glucose cotransporter-2 inhibitors treatment was also asso-
ciated with significant improvement of LAVi in HFrEF patients [MD,
2.19 mL/m
2
; 95% CI (4.26, 0.12); P= 0.04], but was insignificant
in HFpEF patients [MD, 1.46 mL/m
2
; 95% CI (7.13, 4.21);
P= 0.61]. Whereas in other parameters, the only included HFpEF
study was REFORM
20
which included both types of HF but did not
provide the separate results. To identify the effects of SGLT2i on
these parameters in HFrEF patients, we excluded the REFORM
20
study and the remainder of the studies all included HFrEF patients.
The results showed that treatment with SGLT2i significantly
improved the LVESV [SMD, 0.65; 95% CI (1.16, 0.14); P=0.01],
LVESVi [SMD, 0.57; 95% CI (0.94, 0.20); P= 0.003], LVEDV
[SMD, 0.58; 95% CI (0.91, 0.24); P= 0.0007], LVEDVi [SMD,
0.49; 95% CI (0.97, 0.01); P= 0.04] in patients with HFrEF. In
patients with overall HF, the LVEF, LVESVi, and LAVi were significant-
ly improved after SGLT2i therapy, when further limited to HFrEF
patients, the LVEF, LVESVi, LAVi, LVMi, LVESV, LVEDV, and LVEDVi
were significantly improved, which indicated the favourable effect of
SGLT2i was more significant in HFrEF patients (Supplementary ma-
terial online, Table S1).
To identify the effects of SGLT2i on patients with and without
T2DM, corresponding subgroup analyses were performed. Because
Figure 2 Overall meta-analyses of left ventricular systolic function and mass. CI, confidence interval; IV, inverse variance; LVEF, left ventricular ejec-
tion fraction; LVGLS, left ventricular global longitudinal strain; LVM, left ventricular mass; LVMi, left ventricular mass index; SGLT2i, sodium-glucose
cotransporter-2 inhibitor;Std. mean difference, standardized mean difference.
SGLT2 inhibitors and cardiac remodelling 7
Downloaded from https://academic.oup.com/eurjpc/advance-article/doi/10.1093/eurjpc/zwab173/6430917 by guest on 19 November 2021

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
only two studies included patients without or not limited to
T2DM,
25,26
and there was high heterogeneity in the T2DM subgroup
for most parameters. Therefore, subgroup analysis according to
baseline T2DM was only available for LVMi. The result showed that
SGLT2i treatment was associated with significant improvement of
LVMi in patients without or not limited to T2DM [MD, 9.72 g/m
2
;
95% CI (14.19, 5.24); P< 0.0001], but not in T2DM patients [MD,
0.40 g/m
2
; 95% CI (1.46, 0.66); P= 0.46], which might suggest a
more beneficial effect of SGLT2i on non-diabetic heart
(Supplementary material online, Figure S19).
In order to identify which type of SGLT2i might exert a more pre-
dominant effect on reversing cardiac remodelling, we performed a
corresponding subgroup analysis. Treatment with empagliflozin was
associated with significant improvement of LVM [SMD, 0.45; 95%
CI (0.86, 0.05); P= 0.03], LVMi [MD, 4.69 g/m
2
; 95% CI (8.46,
0.93); P= 0.01], LVESV [SMD, 0.53; 95% CI (0.93, 0.13);
Figure 3 Overall meta-analyses of left ventricular volume. CI, confidence interval; IV, inverse variance; LVEDV, left ventricular end-diastolic volume;
LVEDVi, left ventricular end-diastolic volume index; LVESV, left ventricular end-systolic volume; LVESVi, left ventricular end-systolic volume index;
SGLT2i, sodium-glucose cotransporter-2 inhibitor; Std. mean difference, standardized mean difference.
8N. Zhang et al.
Downloaded from https://academic.oup.com/eurjpc/advance-article/doi/10.1093/eurjpc/zwab173/6430917 by guest on 19 November 2021

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
P=0.01], LVESVi [SMD, 0.51; 95% CI (0.88, 0.14); P=0.007],
LVEDV [SMD, 0.43; 95% CI (0.80, 0.06); P= 0.02], and LVEDVi
[SMD, 0.50; 95% CI (0.95, 0.05); P= 0.03] in the overall popula-
tion (Supplementary material online, Table S1), whereas no significant
change was found in the subgroups of dapagliflozin, canagliflozin, and
luseogliflozin. Details about the aforementioned subgroup analyses
and those according to control type, imaging method, and follow-up
duration are presented in Supplementary material online, S3.
Heterogeneity and publication bias
There was high statistical heterogeneity in the overall results of LVEF,
LVESV, LVESVi, LVEDV, LVEDVi, and E/e0, and moderate heterogen-
eity in LVGLS. To identify the source of heterogeneity, sensitivity
analyses by the leave-one-out method were performed. For LVEF,
LVESV, and LVESVi, after removing the EMPA-TROPISM (ATRU-4)
study,
25
heterogeneity largely decreased and without significantly
changing the pooled results. Participants in the EMPA-TROPISM
Figure 4 Overall meta-analyses of left ventricular diastolic function. CI, confidence interval; DT, E-wave deceleration time; E/e0, ratio of early mitral
inflow velocity (E) to early mitral annular velocity (e0); E/A, ratio of early to atrial mitral inflow velocity; Lateral e0, early lateral annular tissue Doppler
velocity; IV, inverse variance; LAVi, left atrial volume index; Septal e0, early septal annular tissue Doppler velocity; SGLT2i, sodium-glucose cotrans-
porter-2 inhibitor; Std. mean difference, standardized mean difference.
SGLT2 inhibitors and cardiac remodelling 9
Downloaded from https://academic.oup.com/eurjpc/advance-article/doi/10.1093/eurjpc/zwab173/6430917 by guest on 19 November 2021

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
(ATRU-4) were all non-T2DM and with HFrEF, which was unique in
the included studies and may contribute to the heterogeneity. After
removing the IDOL-EPOC study
29
which took conventional anti-dia-
betic therapies as control, the heterogeneity of E/e0largely decreased
without changing the pooled result. For LVEDV and LVEDVi, one
study removed sensitivity analyses could not find the source of het-
erogeneity. Other causes of heterogeneity may include SGLT2i dos-
age and geographic variation. Details about the heterogeneity
assessment of individual outcomes are shown in Supplementary ma-
terial online, S3 and results of publication bias assessment are pre-
sented in Supplementary material online, S4.
Discussion
The present meta-analysis evaluated the effects of four types of
SGLT2i in patients with T2DM and/or HF, focusing on cardiac remod-
elling parameters including cardiac function and structure. The main
findings are the following: (i) SGLT2i treatment significantly improved
LV systolic and diastolic function, LV mass and volume, including
LVEF, DT, LVM, LVMi, LVESV, LVESVi, in the overall population; (ii)
The salutary effects of SGLT2i were more significant in patients with
prevalent HF, especially with HFrEF (including improved LVEF, LVMi,
LVESV, LVESVi, LVEDV, LVEDVi, and LAVi in HFrEF patients); (iii)
Among four drugs within the class, empagliflozin has demonstrated a
more predominant role in reversing cardiac remodelling.
Compared with two previously published meta-analyses by Yu
et al.
36
and Patoulias et al.,
37
our study has several strengths. First, our
literature search has been brought up to date in the present study,
thus several recently published, large-scale and high-quality RCTs have
been included.
19,25,26
Second, since echocardiography and CMR rep-
resent the most commonly used and reliable tools to assess cardiac
remodelling,
12
our meta-analysis only included echocardiographic and
CMR studies, whereas those using other methods (e.g. impedance
cardiography) were excluded, which could decrease the potential het-
erogeneity. Third, we provided a detailed description of the statistical
estimation process of the missing means (SDs) and conducted a sensi-
tivity analysis by omitting the imputed studies to testify the robustness
of our results. Fourth, the present study was the first meta-analysis to
analyse the effects of SGLT2i therapy on RV function (TAPSE) and
several LV function parameters (E/A, DT, septal e0, lateral e0).
Previous meta-analyses found that SGLT2i treatment was associ-
ated with significant improvement of LVM and E/e0in the overall
population and LVEF in HFrEF patients, but they failed to demon-
strated a significant improvement of LVMi, LVESV, LVESVi, LVEDV,
LVEDVi, and LAVi.
36,37
Apart from the favourable effects on LVM in
the overall population and LVEF in patients with HFrEF which were
in line with these two meta-analyses,
36,37
our study also demon-
strated the beneficial role of SGLT2i in LVEF, LVMi, LVESV, LVESVi,
and DT in the overall population and also the LVEDV, LVEDVi, and
LAVi in patients with HFrEF. E/e0was also significantly improved in
the overall population after treatment with other three SGLT2i ex-
cept luseogliflozin, which is currently only approved in Japan. The dif-
ferent results aforementioned were mainly because several recently
published, large-scale RCTs were not included in the previous meta-
analyses,
36,37
and Yu et al.
36
also included one study in which the
LVEF was assessed by impedance cardiography. Prior experimental
studies may underpin our results. For example, Santos-Gallego
et al.
16
observed a significant improvement of LVEF and LVM in non-
diabetic porcine model and Lee et al.
38
demonstrated a remarkable
improvement of LVESV and LVEDV in hypertensive HF rats after
treatment with SGLT2i. However, LVGLS, a more sensitive measure-
ment for systolic dysfunction, and TAPSE, the only included param-
eter for RV function, both were not significantly improved after
SGLT2i therapy in our analysis, which may be, at least partially, due to
the limited number of studies and small sample size thus require fur-
ther study.
The underlying mechanisms of cardiac remodelling are complex,
involving molecular events within cells and the interstitium, which to-
gether act to alter the shape, size, and mass of the heart after cardiac
injury. In response to some cardiac injuries, such as pressure and vol-
ume overload, ischaemia/reperfusion, myocardial infarction, and neu-
roendocrine activation, the complex remodelling cascades are
triggered, including inflammation, oxidative stress, metabolic abnor-
malities, mitochondrial dysfunction, autophagy and apoptosis, result-
ing in myocyte loss, cardiac hypertrophy, and interstitial fibrosis.
39,40
Additionally, epigenetic changes including DNA methylation, adeno-
sine triphosphate (ATP)-dependent chromatin remodelling, histone
modifications, and non-coding RNA-related mechanisms are also
considered important factors contributing to adverse cardiac remod-
elling.
41,42
Remodelling is associated with worse prognosis, whereas
its reversal is typically accompanied by improved symptoms, better
quality of life, and lower risk of hospitalization or death.
12
In the pre-
sent study, SGLT2i have been demonstrated to exert remarkably
beneficial effects on reversing cardiac remodelling. Whereas the
mechanisms of the benefits of SGLT2i on cardiac remodelling remain
incompletely understood. Previous experimental and clinical studies
have suggested several potential mechanisms. Firstly, SGLT2i have
been shown to reduce cardiac preload due to the diuretic and natri-
uretic effect, thus could mitigate the LV stretch and wall stress, then
lead to a reduction in LV volume. Secondly, by lowering arterial stiff-
ness and blood pressure, SGLT2i have been associated with a reduc-
tion of cardiac afterload.
19
Thirdly, SGLT2i could reduce cardiac
inflammation via inhibiting activation of the nucleotide-binding do-
main-like receptor protein 3 (NLRP3) inflammasome. Fourthly, re-
cent experimental evidence also suggests that SGLT2i has a
cardioprotective effect against ischaemia/reperfusion injury, poten-
tially through decreasing the calmodulin kinase II activity.
43
Fifthly,
SGLT2i might have downstream epigenetic-oriented effects in car-
diac cells to ameliorate cardiac remodelling.
44
Besides, several add-
itional mechanistic benefits of SGLT2i on the myocardium have been
proposed and might also explain the reversed cardiac remodelling,
including reduction of cardiac oxidative stress, improvement of myo-
cardial energetics and inhibition of the mammalian target of rapamy-
cin pathway.
43
Compared with non-HF patients, those with baseline HF, irre-
spective of glycaemic status, seemed to benefit more from SGLT2i
therapy in our analysis. Previous studies have also suggested that
patients with worse cardiac function or more dilated LV chambers
may have a greater propensity for reverse remodelling,
12
thus HF
patients may have a larger potential for reversing remodelling than
patients with normal baseline cardiac function. After exploring the
effects of SGLT2i on different HF phenotypes, a positive effect on
10 N. Zhang et al.
Downloaded from https://academic.oup.com/eurjpc/advance-article/doi/10.1093/eurjpc/zwab173/6430917 by guest on 19 November 2021

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
HFrEF compared with a neutral effect on HFpEF was observed in our
analysis, consistent with the established different effects of angioten-
sin-converting enzyme inhibitors, angiotensin receptor blockers, and
beta-blockers on these two phenotypes. The discordant outcomes
of similar pharmacological therapy in HFrEF and HFpEF may be
attributed to the different signal transduction cascades of myocardial
remodelling between these two conditions.
45
For example, cellular
proliferation and metabolism-related biomarkers were found to be
specific for HFrEF, whereas inflammation and extracellular matrix re-
organization-related biomarkers were specific for HFpEF.
46
Given
only two studies included HFpEF patients in the present meta-
analysis,further studies are needed to testify our results.
Another novel finding of the present study is the more favourable
effect of empagliflozin (among the four SGLT2i) on improving LVM,
LVMi, LVESV, LVESVi, LVEDV, and LVEDVi. In a previous meta-
analysis of cardiorenal outcomes of SGLT2i, McGuire et al.
47
reported differences in outcomes associated with different SGLT2i,
for example, the beneficial effects on cardiovascular death were only
observed for empagliflozin within the trials. Whether this is due to
differences in the populations and their risk profiles, differences in dis-
ease severity or in the drugs per se requires further exploration.
Pharmacologically, empagliflozin has the highest (2500-fold) select-
ivity for SGLT2 over SGLT1 compared with canagliflozin (250-
fold), dapagliflozin (1200-fold), and luseogliflozin (1650-fold).
48,49
Besides the various selectivity for SGLT2, some potential mecha-
nisms may contribute to their different effects on cardiac remodel-
ling. For instance, empagliflozin could attenuate cardiac fibrosis to
prevent adverse remodelling
38
and switch myocardial fuel utilization
away from glucose towards ketone bodies, free fatty acids, and
branched-chain amino acids, thereby improving myocardial energet-
ics, enhancing LV systolic function and ameliorating adverse LV
remodelling.
16
Shi et al.
17
also observed decreased cardiac remodel-
ling in a mice model after treatment with dapagliflozin. The fact that
empagliflozin was associated with significantly reverse remodelling
might also be related to that more studies have been published with
empagliflozin than other SGLT2i. Therefore, further investigation
into the potential effects of different SGLT2i on cardiac remodelling
and the underlying mechanisms is of utmost importance.
Limitations
First, this meta-analysis included four imputed studies, of which the
results needed to be estimated through several steps, which might in-
fluence the combined results. Given that, we conducted correspond-
ing sensitivity analyses by omitting these studies to testify the
robustness of our results. Second, we included two imaging modal-
ities, CMR and echocardiography. This may lead to between-studies
heterogeneity and exert an impact on the pooled results, in consider-
ation of which, corresponding subgroup analyses were performed.
Third, the number of included studies and patients was limited in the
TAPSE and HFpEF subgroup; therefore, further studies focusing on
the effects of SGLT2i on HFpEF patients are required.
Conclusions
Sodium-glucose cotransporter-2 inhibitors play a substantial role in
reversing adverse cardiac remodelling, including improving LV
systolic and diastolic function, LV mass and volume, especially in
patients with HFrEF. Our results also indicate that empagliflozin is
associated with a greater benefit on cardiac remodelling than other
SGLT2i. This study thus supports a role for SGLT2i in the treatment
of HFrEF patients independently of glycaemic status. The present
findings indicate that reversed cardiac remodelling may partially ex-
plain the favourable effects of SGLT2i on HF.
Supplementary material
Supplementary material is available at European Journal of Preventive
Cardiology online.
Funding
This work was supported by grants from the National Natural
Science Foundation of China (Grant number: 81970270 to T.L.).
Conflict of interest: G.Y.H.L. has been a consultant and speaker
for BMS/Pfizer, Boehringer Ingelheim, and Daiichi-Sankyo. No fees
are received personally. Other authors have no disclosures to
declare.
Data availability
All data relevant to the study are included in the article or uploaded
as a Supplementary material online. No more additional data are
available.
References
1. Zinman B, Wanner C, Lachin JM, Fitchett D, Bluhmki E, Hantel S, Mattheus M,
Devins T, Johansen OE, Woerle HJ, Broedl UC, Inzucchi SE; EMPA-REG
OUTCOME Investigators. Empagliflozin, cardiovascular outcomes, and mortality
in type 2 diabetes. N Engl J Med 2015;373:2117–2128.
2. Neal B, Perkovic V, Mahaffey KW, de Zeeuw D, Fulcher G, Erondu N, Shaw W,
Law G, Desai M, Matthews DR. Canagliflozin and cardiovascular and renal events
in type 2 diabetes. N Engl J Med 2017;377:644–657.
3. Wiviott SD, Raz I, Bonaca MP, Mosenzon O, Kato ET, Cahn A, Silverman MG,
Zelniker TA, Kuder JF, Murphy SA, Bhatt DL, Leiter LA, McGuire DK, Wilding
JPH, Ruff CT, Gause-Nilsson IAM, Fredriksson M, Johansson PA, Langkilde A-M,
Sabatine MS. Dapagliflozin and cardiovascular outcomes in type 2 diabetes. N
Engl J Med 2019;380:347–357.
4. McMurray JJV, Solomon SD, Inzucchi SE, Køber L, Kosiborod MN, Martinez FA,
Ponikowski P, Sabatine MS, Anand IS, B
elohla´ vek J, Bo¨ hm M, Chiang C-E, Chopra
VK, de Boer RA, Desai AS, Diez M, Drozdz J, Duka´t A, Ge J, Howlett JG, Katova
T, Kitakaze M, Ljungman CEA, Merkely B, Nicolau JC, O’Meara E, Petrie MC,
Vinh PN, Schou M, Tereshchenko S, Verma S, Held C, DeMets DL, Docherty
KF, Jhund PS, Bengtsson O, Sjo¨ strand M, Langkilde A-M. Dapagliflozin in patients
with heart failure and reduced ejection fraction. N Engl J Med 2019;381:
1995–2008.
5. Packer M, Anker SD, Butler J, Filippatos G, Pocock SJ, Carson P, Januzzi J, Verma
S, Tsutsui H, Brueckmann M, Jamal W, Kimura K, Schnee J, Zeller C, Cotton D,
Bocchi E, Bo¨ hm M, Choi D-J, Chopra V, Chuquiure E, Giannetti N, Janssens S,
Zhang J, Gonzalez Juanatey JR, Kaul S, Brunner-La Rocca H-P, Merkely B,
Nicholls SJ, Perrone S, Pina I, Ponikowski P, Sattar N, Senni M, Seronde M-F,
Spinar J, Squire I, Taddei S, Wanner C, Zannad F. Cardiovascular and renal out-
comes with empagliflozin in heart failure. N Engl J Med 2020;383:1413–1424.
6. Li D, Liu Y, Hidru TH, Yang X, Wang Y, Chen C, Li KHC, Tang Y, Wei Y, Tse G,
Xia Y. Protective effects of sodium-glucose transporter 2 inhibitors on atrial fib-
rillation and atrial flutter: a systematic review and meta-analysis of randomized
placebo-controlled trials. Front Endocrinol (Lausanne) 2021;12:619586.
7. Ye G, Wang S, Peng D. Effects of SGLT2 inhibitor on ischemic events stemming
from atherosclerotic coronary diseases: a systematic review and meta-analysis
SGLT2 inhibitors and cardiac remodelling 11
Downloaded from https://academic.oup.com/eurjpc/advance-article/doi/10.1093/eurjpc/zwab173/6430917 by guest on 19 November 2021

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
with trial sequential analysis of randomized controlled trials. J Cardiovasc
Pharmacol 2021;77:787–795.
8. Yu B, Dong C, Hu Z, Liu B. Effects of sodium-glucose co-transporter 2 (SGLT2)
inhibitors on renal outcomes in patients with type 2 diabetes mellitus and chron-
ic kidney disease: a protocol for systematic review and meta-analysis. Medicine
(Baltimore) 2021;100:e24655.
9. Benham JL, Booth JE, Sigal RJ, Daskalopoulou SS, Leung AA, Rabi DM. Systematic
review and meta-analysis: SGLT2 inhibitors, blood pressure and cardiovascular
outcomes. Int J Cardiol Heart Vasc 2021;33:100725.
10. Butler J, Usman MS, Khan MS, Greene SJ, Friede T, Vaduganathan M, Filippatos
G, Coats AJS, Anker SD. Efficacy and safety of SGLT2 inhibitors in heart failure:
systematic review and meta-analysis. ESC Heart Fail 2020;7:3298–3309.
11. Bray JJH, Foster-Davies H, Stephens JW. A systematic review examining the
effects of sodium-glucose cotransporter-2 inhibitors (SGLT2is) on biomarkers of
inflammation and oxidative stress. Diabetes Res Clin Pract 2020;168:108368.
12. Aimo A, Gaggin HK, Barison A, Emdin M, Januzzi JL Jr. Imaging, biomarker, and
clinical predictors of cardiac remodeling in heart failure with reduced ejection
fraction. JACC Heart Fail 2019;7:782–794.
13. Khan MS, Felker GM, Pi~
na IL, Camacho A, Bapat D, Ibrahim NE, Maisel AS,
Prescott MF, Ward JH, Solomon SD, Januzzi JL, Butler J. Reverse cardiac remod-
eling following initiation of sacubitril/valsartan in patients with heart failure with
and without diabetes. JACC Heart Fail 2021;9:137–145.
14. Cohn JN, Ferrari R, Sharpe N. Cardiac remodeling—concepts and clinical impli-
cations: a consensus paper from an international forum on cardiac remodeling.
Behalf of an International Forum on Cardiac Remodeling. J Am Coll Cardiol 2000;
35:569–582.
15. Connelly KA, Zhang Y, Visram A, Advani A, Batchu SN, Desjardins J-F, Thai K,
Gilbert RE. Empagliflozin improves diastolic function in a nondiabetic rodent
model of heart failure with preserved ejection fraction. JACC Basic Transl Sci
2019;4:27–37.
16. Santos-Gallego CG, Requena-Ibanez JA, San Antonio R, Ishikawa K, Watanabe S,
Picatoste B, Flores E, Garcia-Ropero A, Sanz J, Hajjar RJ, Fuster V, Badimon JJ.
Empagliflozin ameliorates adverse left ventricular remodeling in nondiabetic heart
failure by enhancing myocardial energetics. J Am Coll Cardiol 2019;73:1931–1944.
17. Shi L, Zhu D, Wang S, Jiang A, Li F. Dapagliflo zin attenuates cardiac remodeling
in mice model of cardiac pressure overload. Am J Hypertens 2019;32:452–459.
18. Shao Q, Meng L, Lee S, Tse G, Gong M, Zhang Z, Zhao J, Zhao Y, Li G, Liu T.
Empagliflozin, a sodium glucose co-transporter-2 inhibitor, alleviates atrial
remodeling and improves mitochondrial function in high-fat diet/streptozotocin-
induced diabetic rats. Cardiovasc Diabetol 2019;18:165.
19. Lee MMY, Brooksbank KJM, Wetherall K, et al. Effect of empagliflozin on left ven-
tricular volumes in patients with type 2 diabetes, or prediabetes, and heart failure
with reduced ejection fraction (SUGAR-DM-HF). Circulation 2021;143:516–525.
20. Singh JSS, Mordi IR, Vickneson K, Fathi A, Donnan PT, Mohan M, Choy AMJ,
Gandy S, George J, Khan F, Pearson ER, Houston JG, Struthers AD, Lang CC.
Dapagliflozin versus placebo on left ventricular remodeling in patients with dia-
betes and heart failure: the REFORM trial. Diabetes Care 2020;43:1356–1359.
21. McGrath S, Zhao X, Steele R, Thombs BD, Benedetti A. Estimating the sample
mean and standard deviation from commonly reported quantiles in meta-ana-
lysis. Stat Methods Med Res 2020;doi:10.1177/0962280219889080.
22. Wiebe N, Vandermeer B, Platt RW, Klassen TP, Moher D, Barrowman NJ. A sys-
tematic review identifies a lack of standardization in methods for handling missing
variance data. J Clin Epidemiol 2006;59:342–353.
23. Eickhoff MK, Olsen FJ, Frimodt-Møller M, Diaz LJ, Faber J, Jensen MT, Rossing P,
Persson F. Effect of dapagliflozin on cardiac function in people with type 2 dia-
betes and albuminuria—a double blind randomized placebo-controlled crossover
trial. J Diabetes Complications 2020;34:107590.
24. Verma S, Mazer CD, Yan AT, Mason T, Garg V, Teoh H, Zuo F, Quan A,
Farkouh ME, Fitchett DH, Goodman SG, Goldenberg RM, Al-Omran M, Gilbert
RE, Bhatt DL, Leiter LA, Ju¨ ni P, Zinman B, Connelly KA. Effect of empagliflozin
on left ventricular mass in patients with type 2 diabetes mellitus and coronary ar-
tery disease: the EMPA-HEART cardiolink-6 randomized clinical trial. Circulation
2019;140:1693–1702.
25. Santos-Gallego CG, Vargas-Delgado AP, Requena-Ibanez JA, Garcia-Ropero A,
Mancini D, Pinney S, Macaluso F, Sartori S, Roque M, Sabatel-Perez F, Rodriguez-
Cordero A, Zafar MU, Fergus I, Atallah-Lajam F, Contreras JP, Varley C, Moreno
PR, Abascal VM, Lala A, Tamler R, Sanz J, Fuster V, Badimon JJ. Randomized trial
of empagliflozin in nondiabetic patients with heart failure and reduced ejection
fraction. J Am Coll Cardiol 2021;77:243–255.
26. Omar M, Jensen J, Ali M, Frederiksen PH, Kistorp C, Videbæk L, Poulsen MK,
Tuxen CD, Mo¨ ller S, Gustafsson F, Køber L, Schou M, Møller JE. Associations of
empagliflozin with left ventricular volumes, mass, and function in patients with
heart failure and reduced ejection fraction: a substudy of the empire HF random-
ized clinical trial. JAMA Cardiol 2021;6:836–840.
27. Brown AJM, Gandy S, McCrimmon R, Houston JG, Struthers AD, Lang CC. A
randomized controlled trial of dapagliflozin on l8eft ventricular hypertrophy in
people with type two diabetes: the DAPA-LVH trial. Eur Heart J 2020;41:
3421–3432.
28. Shim CY, Seo J, Cho I, Lee CJ, Cho I-J, Lhagvasuren P, Kang S-M, Ha J-W, Han G,
Jang Y, Hong G-R. Randomized, controlled trial to evaluate the effect of dapagli-
flozin on left ventricular diastolic function in patients with type 2 diabetes melli-
tus: the IDDIA trial. Circulation 2021;143:510–512.
29. Kayano H, Koba S, Hirano T, Matsui T, Fukuoka H, Tsuijita H, Tsukamoto S,
Hayashi T, Toshida T, Watanabe N, Hamazaki Y, Geshi E, Murakami M,
Aihara K, Kaneko K, Yamada H, Kobayashi Y, Shinke T. Dapagliflozin influen-
ces ventricular hemodynamics and exercise-induced pulmonary hypertension
in type 2 diabetes patients—a randomized controlled trial. Circ J 2020;84:
1807–1817.
30. Oldgren J, Laurila S, A
˚kerblom A, Latva-Rasku A, Rebel os E, Isackson H,
Saarenhovi M, Eriksson O, Heurling K, Johansson E, Wild era¨ng U, Karlsson C,
Esterline R, Ferrannini E, Oscarsson J, Nuutila P. Effects of 6 weeks of treatment
with dapagliflozin, a sodium-glucose co-tra nsporter-2 inhibitor, on myocardial
function and metabolism in patients with type 2 diabetes: a randomize d,
placebo-con trolled, exploratory study. Diabetes Obes Metab 2021;23:
1505–1517.
31. Tanaka A, Hisauchi I, Taguchi I, Sezai A, Toyoda S, Tomiyama H, Sata M, Ueda S,
Oyama J-I, Kitakaze M, Murohara T, Node K; CANDLE Trial Investigators.
Effects of canagliflozin in patients with type 2 diabetes and chronic heart failure:
a randomized trial (CANDLE). ESC Heart Fail 2020;7:1585–1594.
32. Carbone S, Billingsley HE, Canada JM, Bressi E, Rotelli B, Kadariya D, Dixon DL,
Markley R, Trankle CR, Cooke R, Rao K, B Shah K, Medina de Chazal H,
Chiabrando JG, Vecchie´ A, Dell M, Mihalick VL, Bogaev R, Hart L, Van Tassell
BW, Arena R, Celi FS, Abbate A. The effects of canagliflozin compared to sita-
gliptin on cardiorespiratory fitness in type 2 diabetes mellitus and heart failure
with reduced ejection fraction: the CANA-HF study. Diabetes Metab Res Rev
2020;36:e3335.
33. Ejiri K, Miyoshi T, Kihara H, Hata Y, Nagano T, Takaishi A, Toda H, Nanba S,
Nakamura Y, Akagi S, Sakuragi S, Minagawa T, Kawai Y, Nishii N, Fuke S,
Yoshikawa M, Nakamura K, Ito H; MUSCAT-HF Study Investigators. Effect of
luseogliflozin on heart failure with preserved ejection fraction in patients with
diabetes mellitus. J Am Heart Assoc 2020;9:e015103.
34. Ponikowski P, Voors AA, Anker SD, et al. 2016 ESC Guidelines for the diagnosis
and treatment of acute and chronic heart failure: the task force for the diagnosis
and treatment of acute and chronic heart failure of the European Society of
Cardiology (ESC) Developed with the special contribution of the Heart Failure
Association (HFA) of the ESC. Eur Heart J 2016;37:2129–2200.
35. McDonagh TA, Metra M, Adamo M, et al. 2021 ESC Guidelines for the
diagnosis and treatment of acute and chronic heart failure. Eur Heart J 2021;42:
3599–3726.
36. Yu Y-W, Zhao X-M, Wang Y-H, Zhou Q, Huang Y, Zhai M, Zhang J. Effect of
sodium-glucose cotransporter 2 inhibitors on cardiac structure and function in
type 2 diabetes mellitus patients with or without chronic heart failure: a meta-
analysis. Cardiovasc Diabetol 2021;20:25.
37. Patoulias D, Papadopoulos C, Katsimardou A, Kalogirou MS, Doumas M. Meta-
analysis assessing the effect of sodium-glucose co-transporter-2 inhibitors on left
ventricular mass in patients with type 2 diabetes mellitus. Am J Cardiol 2020;134:
149–152.
38. Lee H-C, Shiou Y-L, Jhuo S-J, Chang C-Y, Liu P-L, Jhuang W-J, Dai Z-K, Chen W-
Y, Chen Y-F, Lee A-S. The sodium-glucose co-transporter 2 inhibitor empagliflo-
zin attenuates cardiac fibrosis and improves ventricular hemodynamics in hyper-
tensive heart failure rats. Cardiovasc Diabetol 2019;18:45.
39. Schirone L, Forte M, Palmerio S, Yee D, Nocella C, Angelini F, Pagano F,
Schiavon S, Bordin A, Carrizzo A, Vecchione C, Valenti V, Chimenti I, De Falco
E, Sciarretta S, Frati G. A review of the molecular mechanisms underlying the de-
velopment and progression of cardiac remodeling. Oxid Med Cell Longev 2017;
2017:3920195.
40. Gronda E, Sacchi S, Benincasa G, Vanoli E, Napoli C. Unresolved issues in left
ventricular postischemic remodeling and progression to heart failure. J Cardiovasc
Med (Hagerstown) 2019;20:640–649.
41. Hamdani N, Costantino S, Mu¨gge A, Lebeche D, Tscho¨ pe C, Thum T, Paneni F.
Leveraging clinical epigenetics in heart failure with preserved ejection fraction: a
call for individualized therapies. Eur Heart J 2021;42:1940–1958.
42. Berezin A. Epigenetics in heart failure phenotypes. BBA Clin 2016;6:31–37.
43. Lopaschuk GD, Verma S. Mechanisms of cardiovascular benefits of sodium glu-
cose co-transporter 2 (SGLT2) inhibitors: a state-of-the-art review. JACC Basic
Transl Sci 2020;5:632–644.
44. Napoli C, Bontempo P, Palmieri V, Coscioni E, Maiello C, Donatelli F, Benincasa
G. Epigenetic therapies for heart failure: current insights and future potential.
Vasc Health Risk Manag 2021;17:247–254.
45. Borlaug BA, Paulus WJ. Heart failure with preserved ejection fraction: patho-
physiology, diagnosis, and treatment. Eur Heart J 2011;32:670–679.
12 N. Zhang et al.
Downloaded from https://academic.oup.com/eurjpc/advance-article/doi/10.1093/eurjpc/zwab173/6430917 by guest on 19 November 2021

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
46. Tromp J, Westenbrink BD, Ouwerkerk W, van Veldhuisen DJ, Samani NJ,
Ponikowski P, Metra M, Anker SD, Cleland JG, Dickstein K, Filippatos G, van der
Harst P, Lang CC, Ng LL, Zannad F, Zwinderman AH, Hillege HL, van der Meer
P, Voors AA. Identifying pathophysiological mechanisms in heart failure with
reduced versus preserved ejection fraction. J Am Coll Cardiol 2018;72:1081–1090.
47. McGuire DK, Shih WJ, Cosentino F, Charbonnel B, Cherney DZI, Dagogo-Jack S,
Pratley R, Greenberg M, Wang S, Huyck S, Gantz I, Terra SG, Masiu kiewicz U,
Cannon CP. Association of SGLT2 inhibitors with cardiovascular and kidney outcomes
in patients with type 2 diabetes: a meta-analysis. JAMA Cardiol 2021;6:148–158.
48. Zelniker TA, Braunwald E. Cardiac and renal effects of sodium-glucose co-trans-
porter 2 inhibitors in diabetes: JACC state-of-the-art review. J Am Coll Cardiol
2018;72:1845–1855.
49. Markham A, Elkinson S. Luseogliflozin: first global approval. Drugs 2014;74:
945–950.
SGLT2 inhibitors and cardiac remodelling 13
Downloaded from https://academic.oup.com/eurjpc/advance-article/doi/10.1093/eurjpc/zwab173/6430917 by guest on 19 November 2021