Efficacy and safety of SGLT2 inhibitor empagliflozin in T2DM with established cardiovascular disease: a systematic review and meta-analysis of RCTs

Abstract

Introduction:

Effective treatment approaches are required for type 2 diabetes mellitus (T2DM) and cardiovascular diseases (CVDs) because these two diseases pose a big health issue for humans. Various medications, but with limitations in their efficacy, are available for the treatment of the above-mentioned diseases, but a drug, empagliflozin, a sodium–glucose cotransporter-2 inhibitor, has shown promising benefits for patients with T2DM and CVD, highlighting the need for a thorough review of its efficacy and safety.

Methods:

We searched online databases till 30 June 2024, and a total of 15 trials involving 8296 participants were included in this study. Evaluating the changes in HbA1c was a main concern of our study, while weight-related metrics [such as body weight and Body Mass Index (BMI)], cardiovascular indicators [like systolic blood pressure (SBP), diastolic blood pressure (DBP), changes in left ventricular ejection fraction %, and heart rate], and various laboratory values [including fasting plasma glucose (FPG), C-reactive protein (CRP), hematocrit, estimated glomerular filtration rate] were considered as secondary assessments.

Results:

Our study concluded that empagliflozin significantly reduced HbA1c levels [standardized mean difference (SMD) = −0.62, 95% CI = −0.95 to −0.30, I² = 85%, P = 0.0001], indicating improved glycemic control. Additionally, a reduction in body weight was also noted (SMD = −2.32, 95% CI = −3.42 to −1.21, I² = 54%, P < 0.0001). The drug also demonstrated favorable outcomes for secondary endpoints. Meta-regression and subgroup analyses provided insights into the variability of treatment effects across different populations.

Conclusions:

The efficacy of empagliflozin in managing T2DM and CVD is evaluated by this pooled analysis, highlighting great improvement in glycemic control and weight reduction. Further research is needed for empagliflozin’s long-term safety and efficacy across populations.

Introduction

In the current era, cardiovascular disease (CVD) and type 2 diabetes mellitus (T2DM) are two major health problems faced by the public. These diseases often exist simultaneously, affecting each other’s morbidity and mortality. The patients of T2DM are at increased risk of cardiovascular problems, and this makes it necessary to implement such therapeutic strategies that combine glycemic control with cardiovascular protection^[1]. Among T2DM patients with CVD, to lower blood pressure, ACE inhibitors were proposed as the first line of treatment because ACE inhibitors have a great potential to reduce mortality and the risk of hospitalization in such patients. However, these drugs are associated with side effects, such as cough and angioedema^[2]. The sodium–glucose cotransporter-2 inhibitor (SGLT2i) empagliflozin has exhibited early promise in this regard. Empagliflozin inhibits the reabsorption of glucose from the kidney by blocking SGLT2, which results in glycosuria and hence a reduction in blood glucose level^[3]. Furthermore, empagliflozin has a promising dual benefit therapy because, besides lowering blood glucose levels, it also leads to weight loss, reduced blood pressure, and improved cardiovascular status^[4]. Different randomized controlled trials (RCTs) are evident that empagliflozin is beneficial in reducing hospitalization for heart failure (HF), as well as significant reductions in cardiovascular death and total mortality in T2DM patients^[5].

Prior reviews frequently pooled heterogeneous populations and mixed drug classes, limiting inference specific to empagliflozin and to patients with established CVD. Moreover, clinicians often seek clarity on whether short-term changes translate into consistent directions of effect across doses (10 vs. 25 mg) and whether any early renal signal [estimated glomerular filtration rate (eGFR)] should influence initial prescribing or monitoring strategies. Addressing these practical questions is particularly relevant where cost, access, and patient preferences intersect with safety counseling (e.g., hygiene measures to mitigate genital infections) and early follow-up plans.

However, the relative effectiveness of empagliflozin compared to placebo in individuals with T2DM and established CVD is still actively under investigation. Inconsistent results are reported by currently available studies, so there is an urgent need for a thorough literature review to guide evidence-based practice. Accordingly, our objective was to synthesize RCTs of empagliflozin in adults with T2DM and established CVD to evaluate short-term efficacy (glycemic control, blood pressure, weight) and safety (genital infections, urinary tract infection (UTI), hypoglycemia, discontinuations, serious adverse events), with renal function as a key secondary consideration. This focused approach aims to provide clinicians with concise, high-quality evidence to inform initial prescribing and early monitoring while acknowledging between-study clinical diversity.

Methods

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed by this systematic review and meta-analysis^[6] also, the work has been reported in line with AMSTAR (Assessing the methodological quality of systematic reviews) Guidelines. The literature search was prespecified, and the protocol methods were followed as written; no post hoc analyses were added. In accordance with the TITAN Guidelines 2025 for transparent use of AI in scholarly communication, no AI tools were used in the research design, data collection, analysis, or interpretation; AI assistance was limited solely to language refinement during manuscript preparation^[7].

Data sources and search strategy

A comprehensive literature search was conducted utilizing various online databases, including PubMed, Google Scholar, Scopus, ClinicalTrials.gov, ScienceDirect, and the Cochrane Central Register of Controlled Trials (CENTRAL). The search employed MeSH terms such as “empagliflozin,” “cardiovascular,” and “diabetes,” and was conducted up to 30 June 2024. To identify gray literature, including conference proceedings and presentations, manual searches of journals, websites, and reference lists of pertinent review articles were also performed. Regardless of publication date or sample size, only studies written in English were selected. Detailed search strategies are outlined in Supplemental Digital Content Table S1, available at: http://links.lww.com/MS9/B68.

HIGHLIGHTS.

• Empagliflozin significantly lowers HbA1c in T2DM patients with CVD.

• Notable reductions in weight, BMI, SBP, and DBP observed with empagliflozin.

• Improved FPG, CRP reduction, and hematocrit increase were seen in the treatment group.

• Lower risk of serious adverse events, but increased genital infections noted.

• Consistent efficacy across doses; key predictors include SBP, chronic kidney disease, and diabetes duration.

Data synthesis

EndNote Reference Library (Version X7.5; Clarivate Analytics, Philadelphia, Pennsylvania) was used for duplicate removal and screening. Two reviewers independently conducted initial screenings of titles and abstracts. Then, to ensure adherence to the inclusion criteria, a full review was conducted. Help from a third reviewer was taken to resolve any possible discrepancies. We included RCTs enrolling adults with T2DM and established CVD that compared empagliflozin (10 or 25 mg) versus placebo or usual care. Trials had to report at least one prespecified outcome [HbA1c, fasting plasma glucose (FPG), systolic/diastolic blood pressure (SBP/DBP), or body weight] or a safety outcome (genital infections, UTI, hypoglycemia, serious adverse events, treatment discontinuation). We excluded non-randomized studies, studies without a relevant comparator, pediatric populations, and trials not reporting extractable data. No additional studies were added after the prespecified cutoff.

Data extraction

Two investigators independently performed data extraction using a standardized Microsoft Excel spreadsheet. Changes in HbA1c levels, weight-related measures [body weight, Body Mass Index (BMI)], cardiovascular parameters [SBP, DBP, left ventricular ejection fraction % (LVEF) changes, heart rate], and various laboratory values [FPG, C-reactive protein (CRP), hematocrit, eGFR] were the key outcomes assessed, alongside adverse effects were also documented. SBP, DBP, and the presence of comorbidities [hypertension, myocardial infarction (MI), HF, dyslipidemia, coronary artery disease, history of stroke, peripheral artery disease, chronic kidney disease (CKD), and smoking status] were systematically recorded for each study included in the analysis.

Risk of bias and quality assessment

The quality of 15 RCTs included in our study was assessed using The Cochrane Risk of Bias Tool for Randomized Controlled Trials (RoB-2)^[8], evaluating randomization, deviations from intended interventions, missing outcome data, outcome measurement, and selection bias. Studies were classified as having “low,” “moderate,” or “high” risk of bias. Additionally, crossover RCTs included an extra domain in the quality. Two independent reviewers, in cases where discrepancies arose, they were resolved through consultation with a third reviewer.

Statistical analysis

To conduct all meta-analyses and subgroup analyses, with the dosage of empagliflozin stratified into 10 or 25 mg subgroups, RevMan software (version 5.4.1; Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2020) was used. We calculated the Risk Ratio (RR) for dichotomous outcomes, while we used the standardized mean difference (SMD) to analyze continuous data. All outcomes were reported with 95% confidence intervals (CIs), and a P-value of ≤ 0.05 was considered statistically significant. Publication bias and meta regression analysis were calculated using Comprehensive Meta Analyst (version 3.7). Analyses were performed using a random-effects model, with relative risks and their corresponding 95% CIs being pooled. Statistical significance was determined by a P-value of <0.05 for all outcomes. Meta-regression results were reported as coefficients (Coeff) and two-sided P-values. Heterogeneity was evaluated using the Higgins I² statistic^[7]. Values exceeding I² = 50% indicated substantial heterogeneity, necessitating further investigation through sensitivity analyses utilizing the leave-one-out approach

Results

After a comprehensive literature search that identified 13 020 records, 15 studies (13 RCTs and 2 crossover RCTs) met eligibility (PRISMA, Fig. 1). Outcome data were pooled from 8296 participants (empagliflozin 10 mg or 25 mg: n = 5322; placebo: n = 2974); mean follow-up was 16.6 weeks. The mean follow-up duration across the included studies was 16.6 weeks. The mean ages in the empagliflozin and placebo groups were 63.7 ± 0.11 and 64.2 ± 0.16 years, respectively. Further details regarding the study and baseline characteristics are provided in Tables 1 and 2.

Figure 1.

PRISMA flowchart.

Table 1.

Study characteristics

Author, year	Type of study	Sample size		CVD risk	Empagliflozin dosage	Follow-up, weeks
		Empagliflozin, n	Placebo, n
Adel et al 2022	RCT	45	48	ACS after PCI	10 mg/day	24 weeks
Boer et al 2020	RCT	30	33	Heart Failure	25 mg/day	12 weeks
Cheng et al 2022	RCT	62	62	Hypertension	25 mg/day	12 weeks
Ferdinand et al 2019	RCT	78	72	Hypertension	10–25 mg/day	24 weeks
Gohari et al 2022	RCT	47	48	Coronary artery disease	10 mg/day	28 weeks
Griffin et al 2020	Cross-over RCT	20		Heart failure	10 mg/day	2 weeks (followed by a 2-week washout period and crossover)
Jürgens et al 2021	RCT	45	45	Established CVD	25 mg/day	13 weeks
Mansouri et al 2023	RCT	37	38	Refractory angina	25 mg/day	8 weeks
Mordi et al 2020	Cross-over RCT	23		Congestive heart failure	25 mg/day	6 weeks (followed by a 2-week washout period and crossover)
Prochaska et al 2023	RCT	71	71	Heart failure	10 mg/day	12 weeks
Shimizu et al 2020	RCT	46	50	Acute myocardial infarction	10 mg/day	24 weeks
Tanaka et al 2019	RCT	52	53	Established CVD	10 mg/day	24 weeks
Verma et al 2019	RCT	49	48	Coronary artery disease	10 mg/day	24 weeks
Vernstrøm et al 2024	RCT	30	30	Established CVD	10 mg/day	32 weeks
Zinman et al 2015	RCT	4687	2333	CV risk	10 mg/day and 25 mg/day	4 weeks

n, Number of participants; CVD, cardiovascular disease; ACS, acute coronary syndrome; PCI, percutaneous coronary.

Table 2.

Baseline patient characteristics

Author, year		Adel et al 2022	Boer et al 2020	Cheng et al 2022	Ferdinand et al 2019	Gohari et al 2022	Griffin et al 2020	Jürgens et al 2021	Mansouri et al 2023	Mordi et al 2020	Prochaska et al 2023	Shimizu et al 2020	Tanaka et al 2019	Verma et al 2019	Vernstrøm et al 2024	Zinman et al 2015
Age, years	Empagliflozin	55 (45.5–64)a	68.5 (62–74)a	71.2 ± 4.0	56.5 ± 9.3	62.08 ± 8.02	60 ± 12	66 ± 9	67.46 ± 9.4	69.8 ± 5.7	69.3 ± 7.4	63.9 ± 10.4	65.4 ± 11.1	64 (57–69)a	70 ± 7	63.1 ± 8.6
	Placebo	57 (50–66.75)a	71 (59–74)a	71.7 ± 4.3	57.2 ± 9.3	63.60 ± 7.82		67 ± 9	65.47 ± 7.0		68.5 ± 8.0	64.6 ± 11.6	64.1 ± 9.9	64 (56–72)a	67 ± 7	63.2 ± 8.8
Male sex, n (%)	Empagliflozin	27 (60.0)	20 (66.7)	40 (64.5)	43 (55.1)	23 (48.9)	15 (75)	57 (76)	27 (73.0)	17 (73.9)	60 (84.5)	38 (82.6)	36 (69.2)	44 (90)	22 (73)	3336 (71.2)
	Placebo	29 (60.4)	19 (57.6)	37 (59.7)	36 (50.0)	16 (33.3)		63 (84)	23 (60.5)		62 (87.3)	39 (78.0)	36 (67.9)	46 (96)	24 (80)	1680 (72.0)
Duration of diabetes, years	Empagliflozin	6 (4–8)a	–	–	9.3 ± 6.2	–	–	14 (7.5–19)a	–	8.7 ± 6.3	–	38.3 ± 43.4	13.6 ± 13.2	10 (4–15)a	11.0 (5–19)a	–
	Placebo	6 (2–9)a	–	–	9.3 ± 7.9	–	–	15 (8–18)a	–		–	32.4 ± 43.3	13 ± 8.3	10 (5–15)a	9.5 (5–12)a	–
Weight, kg	Empagliflozin	–	87.4 (76–107.4)a	–	–	–	–	–	–	94.3 ± 17.0	–	70.1 ± 13.7	–	81.5 ± 16.9	–	86.6 ± 19.1
	Placebo	–	87.3 (79–97.2)a	–	–	–	–	–	–		–	68.1 ± 14.4	–	80 ± 18.2	–	86.6 ± 19.1
BMI, kg/m2	Empagliflozin	–	31.2 (28.8–34.7)a	25.5 ± 2.3	36.04 ± 12.83	–	37 ± 7	31.9 ± 5.5	27.03 ± 2.8	33.9 ± 5.6	31.9 ± 4.7	25.2 ± 3.7	26.2 ± 5.1	27.7 ± 4.7	32.2 ± 6.2	30.6 ± 5.3
	Placebo	–	31.3 (28.4–34.2)a	25.8 ± 2.3	35.12 ± 8.29	–		29.8 ± 5.6	28.29 ± 3.1		31.8 ± 4.8	25.2 ± 4.1	26.9 ± 5.5	27.4 ± 5.4	31.9 ± 5.3	30.7 ± 5.2
HbA1c, %	Empagliflozin	–	–	7.7 ± 0.5	8.66 ± 0.11	8.05 ± 0.97	7.1 (6.2–10.5)a	7.7 ± 1.0	7.43 ± 0.58	7.9 ± 3.8	7.3 (7.0–7.7)a	6.82 ± 1.00	7.2 ± 0.8	7.9 ± 0.80	7.4 (6.9–7.9)a	8.07 ± 0.85
	Placebo	–	–	7.7 ± 1.1	8.51 ± 0.13	7.75 ± 0.94		7.5 ± 0.9	7.42 ± 0.59		7.4 (7.0–8.2)a	6.89 ± 0.92	7.2 ± 0.9	8 ± 0.9	7.5 (6.8–8.4)a	8.08 ± 0.84
FPG, mg/dL	Empagliflozin	–	–	151.2 ± 14.4	177.7 ± 55.9	163.14 ± 40.84	–	–	–	–	–	–	–	–	–	–
	Placebo	–	–	154.8 ± 18.0 mmol	175.0 ± 46.3	167.74 ± 48.64	–	–	–	–	–	–	–	–	–	–
eGFR, mL/min	Empagliflozin	–	63.5 (56.9–73.1)a	83.0 ± 7.0	91.15 ± 18.95	–	69.1 ± 19.0	79.5 ± 24.0	–	–	–	64.6 ± 15.0	67 ± 12.5	88 ± 16	84 ± 19	74.2 ± 21.6
	Placebo	–	66.5 (55.5–78.8)a	81.5 ± 8.6	91.49 ± 20.79	–		79.3 ± 21.9	–	–	–	66.1 ± 15.7	69.2 ± 13.9	88 ± 18	91 ± 22	73.8 ± 21.1
SBP, mm of Hg	Empagliflozin	–	131 (120.7–139.7)a	153.0 ± 36.8	–	–	126 ± 18	139 ± 16	146.27 ± 25.41	125.2 ± 18.4	134.5 ± 15.9	129.7 ± 11.9	132.8 ± 15.2	139 ± 15	132 ± 11	135.3 ± 16.9
	Placebo	–	128 (119.3–134.7)a	150.8 ± 37.2	–	–		140 ± 13	133.95 ± 30.84		132.8 ± 15.9	123.1 ± 15.7	133 ± 14.5	138 ± 15	134 ± 15	135.8 ± 17.2
DBP, mm of Hg	Empagliflozin	–	76.3 (73–80.3)a	96.2 ± 23.4	–	–	–	81 ± 11	85.24 ± 13.65	70.2 ± 8.9	77.4 ± 9.2	–	76.4 ± 11.5	80 ± 9	77 ± 8	76.6 ± 9.7
	Placebo	–	73.3 (64.7–79)a	95.4 ± 22.3	–	–	–	81 ± 10	83.08 ± 15.59		76.1 ± 9.3	–	74.9 ± 9.5	78 ± 7	82 ± 7	76.8 ± 10.1
Hypertension, n (%)	Empagliflozin	26 (57.8)	–	14 (22.5)	–	44 (93.6)	19 (95)	–	14 (37.8)	–	60 (84.5)	38 (82.6)	41 (78.8)	45 (92)	–	–
	Placebo	32 (66.7)	–	17 (27.4)	–	43 (89.6)		–	14 (36.8)	–	67 (94.4)	39 (78.0)	36 (67.9)	43 (90)	–	–
Myocardial infarction, n (%)	Empagliflozin	–	4 (13.3)	–	–	–	–	–	27 (73)	–	16 (23.2)	–	12 (23.1)	–	10 (33)	–
	Placebo	–	4 (12.1)	–	–	–	–	–	22 (59.5)	–	26 (37.1)	–	13 (24.5)	–	8 (27)	–
Heart failure, n (%)	Empagliflozin	–	–	–	–	–	–	–	12 (32.4)	–	11 (15.5)	–	23 (44.2)	2 (4)	1 (3)	462 (9.9)
	Placebo	–	–	–	–	–	–	–	13 (34.2)	–	19 (26.8)	–	19 (35.8)	4 (8)	2 (7)	244 (10.5)
Dyslipidemia, n (%)	Empagliflozin	–	–	–	–	38 (80.9)	16 (80)	–	–	–	60 (84.5)	34 (73.9)	39 (75.0)	–	–	–
	Placebo	–	–	–	–	34 (70.8)		–	–	–	64 (90.1)	36 (72.0)	38 (71.7)	–	–	–
Coronary artery disease, n (%)	Empagliflozin	–	6 (20)	–	6 (7.7)	–	12 (60)	–	–	10 (43.5)	23 (35.9)	–	–	–	–	3545 (75.6)
	Placebo	–	4 (12.1)	–	5 (6.9)	–		–	–		30 (44.1)	–	–	–	–	1763 (75.6)
Peripheral artery disease, n (%)	Empagliflozin	–	1 (3.3)	–	1 (1.3)	–	–	3 (7)	–	–	7 (10.1)	–	–	2 (4)	–	982 (21.0)
	Placebo	–	1 (3)	–	2 (2.8)	–	–	3 (7)	–	–	8 (11.9)	–	–	3 (6)	–	479 (20.5)
History of stroke, n (%)	Empagliflozin	–	–	–	–	–	–	6 (13)	2 (5.4)	–	5 (7.1)	–	–	8 (16)	5 (17)	1084 (23.1)
	Placebo	–	–	–	–	–	–	6 (13)	3 (7.9)	–	7 (9.6)	–	–	6 (13)	2 (7)	553 (23.7)
Chronic kidney disease, n (%)	Empagliflozin	4 (8.9)	2 (6.7)	–	–	–	–	–	–	–	9 (13.0)	–	–	–	3 (10)	–
	Placebo	3 (6.3)	2 (6.1)	–	–	–	–	–	–	–	14 (20.0)	–	–	–	3 (10)	–
Smokers, n (%)	Empagliflozin	9 (20)	1 (3.3)	–	–	–	–	–	7 (18.9)	–	12.7% (9)	24 (52.2)	9 (17.3)	20 (41)	6 (20)	–
	Placebo	8 (16.7)	4 (12.1)	–	–	–	–	–	9 (23.7)	–	15.5% (11)	27 (54.0)	13 (24.5)	22 (46)	6 (20)	–

Data are presented as n (%) or mean ± standard deviation.

^aData are given as median (interquartile range).

Risk of bias

Low risk of bias is generally shown by all the studies included in this meta-analysis. However, some problems were noted. The RCT done by Zinman et al had issues with the randomization. Also, Adel et al and Griffin et al’s crossover RCT had gaps in reporting outcome data, which might have influenced the results and hence found potential bias. Prochaska et al and Ferdinand et al raised concerns related to outcome measurement. Finally, Tanaka et al exhibited bias in the selection of reported results. A detailed individual assessment of bias risks is provided in Supplemental Digital Content Figures S1 and S2, available at: http://links.lww.com/MS9/B68.

Primary outcome

Change in HbA1c%

Changes in HbA1c%, which was our primary outcome, were reported in 12 studies. Pooling the data revealed that in comparison to the patients receiving placebo, the patients who received empagliflozin showed a great decrease in HbA1c% levels. However, significant heterogeneity between studies was observed (SMD = −0.62, 95% CI = −0.95 to −0.30, I² = 85%, P = 0.0001). The forest plot depicting the change in HbA1c% is presented in Figure 2.

Figure 2.

Change in HbA1c% forest plot.

Weight-related outcomes

A notable decrease in body mass was demonstrated by patients receiving empagliflozin relative to those administered with placebo, as evidenced by an examination of data extracted from eight studies. Moderate heterogeneity in the results was observed (SMD = −2.32, 95% CI = −3.42 to −1.21, I² = 54%, P < 0.0001). Additionally, when data from three studies were pooled, a significant decline in BMI was seen, with no interstudy heterogeneity (SMD = −0.66, 95% CI = −0.87 to −0.44, I² = 0%, P < 0.00001). Forest plots illustrating these weight-related outcomes are presented in Figure 3A–B.

Figure 3.

(A) Change in body weight (kg) forest plot. (B) Change in body mass index (kg/m²) forest plot.

Cardiovascular parameters

Patients who were given empagliflozin showed a significant decrease in SBP (SMD = −0.64, 95% CI = −1.13 to −0.14, I² = 93%, P = 0.01), as did DBP (SMD = −1.16, 95% CI = −2.14 to −0.17, I² = 97%, P = 0.02). Both results achieved statistical significance. However, for both blood pressure parameters, the included studies showed a notable heterogeneity. LVEF showed no significant change when data from four studies were pooled and analyzed (SMD = 0.01, 95% CI = −0.20 to 0.23, I² = 0%, P = 0.91). Additionally, no significant changes in heart rate (bpm) were observed in a cumulative analysis of six trials (MD = −0.02, 95% CI = −1.74 to 1.69, I² = 0%, P = 0.98). Forest plots illustrating these cardiovascular parameters are presented in Figure 4A–D.

Figure 4.

(A) Change in systolic blood pressure (mm of Hg) forest plot. (B) Change in diastolic blood pressure (mm of Hg) forest plot. (C) Change in left ventricular ejection fraction % forest plot. (D) Change in heart rate (beats/min) forest plot.

Laboratory values

In four studies, FPG was analyzed, and it was statistically significantly decreased in the group of patients receiving empagliflozin as compared to the group receiving placebo (MD = −18.02, 95% CI = −23.61 to −12.43, I² = 0%, P < 0.00001). Similarly, CRP levels significantly decreased among patients receiving empagliflozin in three studies (SMD = −0.34, 95% CI = −0.59 to −0.08, I² = 0%, P = 0.009). Moreover, the group receiving empagliflozin showed significant increases in hematocrit levels, albeit with moderate interstudy heterogeneity (SMD = 0.73, 95% CI = 0.62 to 0.84, I² = 59%, P < 0.00001). Alternatively, nine studies evaluated eGFR, and it was decreased in patients receiving empagliflozin as compared to the patients receiving placebo, but this was a nonsignificant difference, and a moderate level of heterogeneity was noted (SMD = −0.08, 95% CI = −0.81 to 0.01, I² = 50%, P = 0.10). Forest plots depicting these outcomes are presented in Figure 5A–D.

Figure 5.

(A) Change in fasting plasma glucose (mg/dL) forest plot. (B) Change in C-reactive protein (mg/L) forest plot. (C) Change in hematocrit % forest plot. (D) Change in estimated glomerular filtration rate (mL/min) forest plot.

Adverse events

The analysis showed that there is a similar risk of experiencing any adverse events among patients receiving empagliflozin (RR = 0.98, 95% CI = 0.97 to 1.00, I² = 0%, P = 0.04) and of discontinuing treatment due to adverse events (RR = 0.89, 95% CI = 0.81 to 0.99, I² = 0%, P = 0.03). Moreover, there was a significant decrease in the risk of multiple serious adverse events in the empagliflozin group relative to the placebo (RR = 0.90, 95% CI = 0.85 to 0.96, I² = 0%, P = 0.0006). Across three studies, a slightly elevated risk of hypoglycemia was observed in the group receiving empagliflozin, but this was not statistically significant and showed slight heterogeneity (RR = 1.07, 95% CI = 0.76 to 1.50, I² = 12%, P = 0.70). Conversely, a statistically significant three-fold increased risk of developing genital infections was observed in patients receiving empagliflozin (RR = 3.61, 95% CI = 2.65 to 4.91, I² = 0%, P < 0.00001). No significant risk of UTIs was identified (RR = 0.99, 95% CI = 0.89 to 1.10, I² = 0%, P = 0.91). Forest plots illustrating these adverse events are presented in Figure 6A–F.

Figure 6.

(A) ≥1 Adverse event forest plot. (B) ≥1 Adverse event causing discontinuation forest plot. (C) ≥1 Serious adverse event causing discontinuation. (D) Hypoglycemia forest plot. (E) Urinary tract infection forest plot. (F) Genital infection.

Subgroup analysis

The studies were subgrouped based on the dosage of empagliflozin, either 10 or 25 mg, where possible. Overall, there was a significant reduction in HbA1c% levels in the group treated with empagliflozin, and the same results occurred when studies were analyzed according to dosage subgroup. However, significant heterogeneity was observed in each subgroup (overall: SMD = −0.48, 95% CI = −0.83 to −0.12, I² = 80%, P = 0.008; 10 mg subgroup: SMD = −0.88, 95% CI = −1.70 to −0.06, I² = 91%, P = 0.03). Furthermore, there were no significant differences between subgroups (I² = 0%).

With the findings that were consistent across subgroup analysis, there was an overall decline in the weight of the patients. However, variations are seen in heterogeneity; it was absent in the group that received a 10 mg dosage but was notable in the 25 mg subgroup: SMD = −2.65, 95% CI = −4.50 to −0.80, I² = 74%, P = 0.005. When comparing the subgroups in the analysis, no significant differences were found in the outcomes (I² = 0%).

FPG decreased significantly overall with empagliflozin (MD −18.29, 95% CI −23.93 to −12.66, I² = 0%, P < 0.00001). In dose-stratified analyses, the 10 mg subgroup estimate was imprecise and not significant (MD −1.98, 95% CI −45.31 to 45.35, P = 0.93), and the 25 mg subgroup showed a numerically smaller, nonsignificant reduction. No between-subgroup difference was detected. No significant differences were found between subgroups (I² = 0%). Similarly, in the 10 mg subgroup, an overall significant decrease in CRP levels was observed but not in the 25 mg subgroup (10 mg subgroup: SMD = −0.32, 95% CI = −0.58 to −0.05, I² = 0%, P = 0.02; 25 mg subgroup: SMD = −0.57, 95% CI = −1.45 to 0.31, P = 0.20). No heterogeneity or subgroup differences were noted (I² = 0%). With moderate overall heterogeneity, there was a significant increase in hematocrit levels in patients receiving empagliflozin. Both subgroups followed this trend, with heterogeneity decreasing to mild in the 25 mg subgroup, and no subgroup differences were observed (overall: SMD = 0.65, 95% CI = 0.44 to 0.87, I² = 64%, P < 0.00001; 25 mg subgroup: SMD = 0.74, 95% CI = 0.46 to 1.02, I² = 47%, P = 0.20, I² = 0%). An overall nonsignificant decline with moderate heterogeneity was seen in the eGFR in the 10 mg subgroup. However, a numerical nonsignificant decline was found in the 25 mg subgroup, with no heterogeneity observed among studies (overall: SMD = −0.12, 95% CI = −0.34 to 0.10, I² = 64%, P = 0.28; 25 mg subgroup: SMD = −0.08, 95% CI = −0.14 to 0.02, P = 0.005). There were no significant subgroup differences (I² = 0%).

An overall significant decrease in SBP was seen in the subgroup of patients receiving 10 mg empagliflozin, and no heterogeneity was detected. In contrast to this, a nonsignificant decrease but with significant heterogeneity among studies was seen in the 25 mg empagliflozin subgroup (SMD = −0.26, 95% CI = −0.43 to −0.09, I² = 0%, P = 0.002; SMD = −1.08, 95% CI = −2.54 to 0.37, I² = 97%, P = 0.14). Slight subgroup differences were noticed (I² = 17.5%). Identical trends were seen in changes in DBP (SMD = −0.33, 95% CI = −0.52 to −0.14, I² = 0%, P = 0.0007; SMD = −2.43, 95% CI = −6.23 to 1.36, I² = 99%, P = 0.21) (I² = 17.5%). In the subgroup of 10 mg empagliflozin, a nonsignificant decrease was seen. However, a nonsignificant increase was noted in the 25 mg subgroup. No heterogeneity and no subgroup differences were seen (SMD = −0.04, 95% CI = −2.01 to 1.92, I² = 0%, P = 0.97; SMD = −0.03, 95% CI = −3.48 to 3.55, I² = 0%, P = 0.98) (I² = 0%).

Patients receiving empagliflozin showed an overall significant increase in the risk of genital infections and these trends continued subgrouping (RR = 3.56, 95% CI = 2.55 to 4.97, I² = 0%, P <0.00001; RR = 3.62, 95% CI = 2.60 to 5.04, I² = 0%, P <0.00001) with no subgroup differences being apparent (I² = 0%).

Forest Plots for all subgroup analyses are presented in Supplemental Digital Content Figures S3–S12, available at: http://links.lww.com/MS9/B68.

Sensitivity analysis

A sensitivity analysis was conducted for all the initial analyses, and significant heterogeneity was noted (I² > 75%). This involved a leave-one-out analysis, where each study was sequentially removed to assess its impact on heterogeneity. By excluding the study by Cheng et al, a substantial decrease in heterogeneity was observed for both the outcomes: change in SBP (I² = 93% to I² = 38%) and change in DBP (I² = 97% to I² = 1%). Forest plots illustrating the leave-one-out analyses are provided in Supplemental Digital Content Figures S13–S14, available at: http://links.lww.com/MS9/B68.

Publication bias

Publication bias was assessed for all outcomes with 10 or more studies using Egger’s and Begg’s tests. The outcomes evaluated included change in HbA1c% (Egger’s test P-value: 0.83684; Begg’s test P-value: 0.73170), change in SBP (Egger’s test P-value: 0.30908; Begg’s test P-value: 0.58329), and change in hematocrit (Egger’s test P-value: 0.23960; Begg’s test P-value: 0.47427). No evidence of publication bias was found from these tests. Funnel plots for each outcome can be found in Supplemental Digital Content Figures S15–S17, available at: http://links.lww.com/MS9/B68.

Meta-regression

In our primary outcome analysis of change in HbA1c%, various covariates, including mean age, male gender, duration of diabetes, hypertension, prior MI, HF, dyslipidemia, peripheral arterial disease, CKD, stroke/transient ischemic attack, smoking status, and baseline BMI, eGFR, SBP, and DBP, were assessed by performing a meta-regression. Except for the duration of diabetes, insignificant two-sided P-values were shown by all the covariates (Coeff: −0.0751, P = 0.0203), baseline SBP (Coeff: −0.0527, P = 0.0075), and CKD (Coeff: −0.0547, P = 0.0397), which were statistically significant. Detailed results of all meta-regression analyses are presented in Supplemental Digital Content Table S2, available at: http://links.lww.com/MS9/B68, and scatter plots are included in Supplemental Digital Content Figures S18–S32, available at: http://links.lww.com/MS9/B68.

Discussion

The effectiveness of empagliflozin relative to placebo in patients with CVD and T2DM is evaluated in this meta-analysis. In patients receiving empagliflozin, the pooled data from 13 RCTs and 2 crossover RCTs reveal significant improvements in glycemic control, weight reduction, blood pressure management, and important laboratory values. Though the risk of genital infections was increased, empagliflozin was not associated with higher overall adverse events. The primary endpoint was to measure the changes in HbA1c levels, and a noticeable reduction was seen. Compared to placebo, Empagliflozin shows that in patients with T2DM, SGLT2i effectively lowers blood glucose levels, consistent with previous studies^[9–12]. Consistent with the EMPA-REG trial, impaired renal function was associated with a smaller decrease in HbA1c with empagliflozin^[13–15]. This reduction in HbA1c levels is important as it highlights the role of empagliflozin in long-term management of diabetes, particularly in preventing complications associated with poor glycemic control. This highlights the role of empagliflozin in long-term management of diabetes, particularly in preventing complications associated with poor glycemic control. Furthermore, subgroup analysis revealed that there is a consistent reduction in HbA1c levels for both 10 and 25 mg doses, which highlights the efficacy of empagliflozin across different dosing regimens

Similar findings are also revealed by other meta-analyses and systematic reviews, for example considerable impact of empagliflozin and other SGLT2 inhibitors on glycemic control is demonstrated by Liakos et al^[16] and Zaccardi et al^[17]. Just like our results, it was also revealed by Liakos et al that, as compared to placebo, HbA1c levels are significantly reduced by empagliflozin. The class-wide glycemic benefits observed by Zaccardi et al are reflected here with empagliflozin specifically. However, our study also suggests that patient characteristics such as baseline HbA1c, duration of diabetes, baseline SBP, and CKD influence outcomes; therefore, it adds to the literature by providing a more detailed analysis of heterogeneity and its potential sources.

Empagliflozin is of dual benefit in promoting body-weight loss in addition to glycemic control, as it reduces body weight and BMI^[18,19]. However, unintentional weight loss due to empagliflozin in patients with HF and reduced ejection fraction is linked with higher mortality risk, known as the obesity paradox^[20–22]. A systematic meta-analysis involving 14 737 patients with HF demonstrated that empagliflozin’s favorable impacts on cardiovascular hospitalization and mortality due to HF were consistent across the BMI categories, implying that BMI does not significantly alter treatment outcomes^[23].

The results highlight that empagliflozin is of great benefit in weight reduction, but one should be careful while considering its use in patients with HF, as weight loss might have different implications in HF patients. Moreover, empagliflozin lowers both systolic and diastolic pressure effectively, although high heterogeneity (I²) was noted. A prespecified leave-one-out sensitivity analysis identified the study by Cheng et al as a key driver of blood-pressure heterogeneity. Excluding this study reduced I² for SBP from 93 to 38% and for DBP from 97 to 1%, without changing the direction of effect. This might be due to differences in patient demographics, such as baseline HbA1c levels, diabetes duration, and concurrent medications. These findings strengthen confidence in the pooled BP results. The mechanisms are likely related to reductions in intravascular volume and alterations in renal hemodynamics, which contribute to improvements in blood pressure, slight weight loss, and decreased albuminuria^[18]. Empagliflozin has also been found to trigger transient natriuresis, redistributing sodium delivery from the proximal tubule to distal segments and activating tubuloglomerular feedback, which reduces intraglomerular pressure and may optimize heart–kidney interactions^[24,25]. It is notable that the cardiovascular benefits of empagliflozin are more consistent with hemodynamic effects rather than direct cardiac effects, as no significant changes in LVEF or heart rate were observed in our study. This is supported by evidence that the drug reduces arterial stiffness and cardiac workload, most likely due to reduced blood pressure from decreased plasma volume^[26].

Previous studies, such as those by Neal et al^[27] in the CANVAS program, it was revealed that blood pressure and body weight are consistently reduced by SGLT2 inhibitors, including empagliflozin. A study DECLARE-TIMI 58 trial^[28], which demonstrated similar outcomes with dapagliflozin, another SGLT2i, further supports the blood pressure-lowering effect of empagliflozin, as observed in our study. Our results extend this by detailing SBP and DBP changes within an established CVD cohort.

Empagliflozin effectively reduces FPG and CRP levels while increasing hematocrit. The decrease in FPG is also supported by previous studies^[11,12,29]. These results indicate improved glycemic control, potential anti-inflammatory effects, and hemoconcentration/erythropoiesis benefits. Short-term eGFR change was overall not significant; a slight numerical decline of 25 mg in subgroup analyses should be interpreted cautiously due to heterogeneity and limited follow-up. Meanwhile, the decrease in CRP points to possible anti-inflammatory properties, which may contribute to the cardiovascular advantages^[30]. The rise in hematocrit reflects the diuretic effects of SGLT2 inhibitors, leading to hemoconcentration and improved erythropoiesis^[31]. Supported by data from several phase I–III trials and pooled analyses, it is notable that Empagliflozin has a well-documented safety profile^[32–36]. Our analysis showed a lower or similar risk of adverse events, discontinuation due to adverse events, and serious adverse events compared to placebo. Nevertheless, there was an increased risk of genital infections, a known adverse event associated with SGLT2 inhibitors^[15,37]. Though some literature suggests a potential risk for UTIs with SGLT2 inhibitors, the risks of hypoglycemia and UTIs were not significantly different from placebo in our pooled analysis^[38–40]. Nonetheless, further investigations are needed, as conflicting data exist in the literature, including meta-analyses and FDA warnings about serious UTI risks associated with SGLT2 inhibitors^[41–43]. Given mixed external reports, clinicians should counsel on genital-infection hygiene and monitor as clinically indicated.

The safety profile results observed in our analysis are consistent with findings from the EMPA-REG OUTCOME trial^[12–14], which reported a lower incidence of severe adverse events and treatment discontinuations linked to empagliflozin. Our study corroborates these safety findings and adds value by providing a comprehensive evaluation of the risks associated with different adverse events, highlighting the need for careful monitoring, especially concerning genital infections.

Subgroup analyses supported the overall findings, with reductions in HbA1c, body weight, and blood pressure across different doses. Emphasizing the importance of individualized treatment approaches, meta-regression highlighted duration of diabetes, baseline SBP, and CKD as significant covariates for HbA1c change. These results highlight the value of personalized approaches in managing CVD and T2DM. Strengths of this study include an extensive literature review, thorough bias assessment, and advanced statistical analysis. The inclusion of RCTs increases the reliability of the conclusions, while the use of the Cochrane Risk of Bias Tool ensures high methodological quality. However, significant heterogeneity in some outcomes, short mean follow-up duration (16.6 weeks), and inclusion of only English-language studies are limitations that may affect the generalizability of the findings. Our results advocate for empagliflozin as an effective therapeutic option for patients with CVD and T2DM, demonstrating significant benefits in glucose control, weight reduction, and blood pressure management, alongside a well-established safety profile. Despite these advantages, the adoption of SGLT2 inhibitors remains below expectations, likely due to issues such as limited insurance coverage and high patient costs, as well as insufficient awareness of these agents across medical specialties^[44,45]. Healthcare providers should consider empagliflozin for individuals with poorly controlled diabetes and increased cardiovascular risk, with counseling about genital-infection risk and early follow-up. Future research should prioritize longer follow-up and standardized adverse-event reporting to clarify renal trajectories and the durability of benefit.

Conclusion

Empagliflozin demonstrates significant benefits in glycemic control, weight reduction, blood pressure management, and key laboratory markers, with a favorable safety profile compared to placebo in patients with CVD and T2DM. Despite some variability in study outcomes, the overall evidence supports the therapeutic use of empagliflozin in this patient population. Further research is essential to elucidate its long-term effects and optimize both safety and efficacy profiles.

Ethical approval

Ethics approval was not required for this systematic review.

Consent

Informed consent was not required for this systematic review.

Sources of funding

No funding was provided for research.

Author contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis, and interpretation, or in all these areas; took part in drafting, revising, or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Conflicts of interest disclosure

There was no conflict of interest among the authors.

Guarantor

Salih Abdella Yusuf.

Research registration unique identifying number (UIN)

It is registered on Prospero CRD42024609977. Status: Review Completed. URL: https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=610029.

Provenance and peer review

No, it was not invited.

Data availability statement

The dataset supporting the conclusions of this article is included in this article.