Skip main navigation

Potential Mediators for Treatment Effects of Novel Diabetes Medications on Cardiovascular and Renal Outcomes: A Meta‐Regression Analysis

Originally publishedhttps://doi.org/10.1161/JAHA.123.032463Journal of the American Heart Association. 2024;13:e032463

Abstract

Background

Prior research suggests clinical effects of glucagon‐like peptide‐1 receptor agonists (GLP‐1RAs) and sodium‐glucose cotransporter‐2 inhibitors (SGLT2is) are mediated by changes in glycated hemoglobin, body weight, systolic blood pressure, hematocrit, and urine albumin‐creatinine ratio. We aimed to confirm these findings using a meta‐analytic approach.

Methods and Results

We updated a systematic review of 9 GLP‐1RA and 13 SGLT2i trials and summarized longitudinal mediator data. We obtained hazard ratios (HRs) for cardiovascular, renal, and mortality outcomes. We performed linear mixed‐effects modeling of LogHRs versus changes in potential mediators and investigated differences in meta‐regression associations among drug classes using interaction terms. HRs generally became more protective with greater glycated hemoglobin reduction among GLP‐1RA trials, with average HR improvements of 20% to 30%, reaching statistical significance for major adverse cardiovascular events (ΔHR, 23%; P=0.02). Among SGLT2i trials, associations with HRs were not significant and differed from GLP1‐RA trials for major adverse cardiovascular events (Pinteraction=0.04). HRs for major adverse cardiovascular events, myocardial infarction, and stroke became less efficacious (ΔHR, −15% to −34%), with more weight loss for SGLT2i but not for GLP‐1RA trials (ΔHR, 4%−7%; Pinteraction<0.05). Among 5 SGLT2i trials with available data, HRs for stroke became less efficacious with larger increases in hematocrit (ΔHR, 123%; P=0.09). No changes in HRs by systolic blood pressure (ΔHR, −11% to 9%) and urine albumin‐creatinine ratio (ΔHR, −1% to 4%) were found for any outcome.

Conclusions

We confirmed increased efficacy findings for major adverse cardiovascular events with reduction in glycated hemoglobin for GLP1‐RAs. Further research is needed on the potential loss of cardiovascular benefits with increased weight loss and hematocrit for SGLT2i.

Nonstandard Abbreviations and Acronyms

GLP‐1RA

glucagon‐like peptide‐1 receptor agonist

MACE

major adverse cardiovascular event

SGLT2i

sodium‐glucose cotransporter‐2 inhibitors

uACR

urine albumin‐creatinine ratio

Clinical Perspective

What Is New?

  • Meta‐regression of 9 GLP‐1RA (glucagon‐like peptide‐1 receptor agonist) and 13 SGLT2i (sodium‐glucose cotransporter‐2 inhibitor) trials showed treatment effects for major adverse cardiovascular events generally became more protective with greater glycated hemoglobin reduction for GLP‐1RAs, but not for SGLT2i.

  • Treatment efficacy for major adverse cardiovascular events, myocardial infarction, and stroke tended to diminish with larger body weight loss for SGLT2is, but not for GLP1‐RAs.

  • Limited data from SGLT2i trials indicated loss of treatment effect for stroke with greater increase in hematocrit.

What Are the Clinical Implications?

  • When treating patients with type 2 diabetes with GLP1‐RAs, more significant cardiovascular risk reduction could be expected when better glycemic control is achieved.

  • Caution may be advised when SGLT2is increase body weight or hematocrit in patients at high atherothrombotic risk, warranting further research based on individual participant data.

Recent cardiovascular and renal outcome trials have demonstrated that the glucose‐lowering drug classes, GLP‐1RAs (glucagon‐like peptide‐1 receptor agonists) and SGLT2i (sodium‐glucose cotransporter‐2 inhibitors), not only improve glycemic control but also decrease adverse atherosclerotic, heart failure, and renal events.1, 2, 3 Relative benefits for atherosclerotic cardiovascular outcomes including stroke are expected to be more pronounced with GLP‐1RAs, whereas relative reductions of hospitalizations for heart failure and progression of kidney disease are expected to be more pronounced with SGLT2i. These findings of clinical effectiveness have led to major changes in clinical practice guidelines for the prevention and risk management of cardiovascular and renal outcomes in patients with diabetes.4, 5

Small experimental studies, as well as secondary and post hoc analyses of aforementioned large clinical outcome trials, have been conducted to better understand the potential mechanisms of action underlying the cardiovascular and renal outcome benefits with GLP‐1RAs and SGLT2i. In these studies, hypotheses of different potential mediation mechanisms have been investigated including improved glycemic control (glycated hemoglobin [HbA1c]),6 improvements in traditional cardiovascular risk factor levels (eg, blood pressure),7 improved nutrient deprivation signaling (with increased erythropoiesis as an important marker),8 and improved low‐grade inflammation from obesity through body weight loss.9 Participant‐level mediation analyses of GLP‐1RA trials identified reductions in HbA1c, urine albumin‐creatinine ratio (uACR), and systolic blood pressure as potential mediators for effects on major adverse cardiovascular events (MACE) and renal outcomes.10, 11, 12 For SGLT2i, reductions in body weight, systolic blood pressure, and uACR, as well as increases of hematocrit, were identified as potential mediators for MACE, heart failure, and renal outcomes,13, 14, 15, 16 whereas improvement in HbA1c is considered a potential mediator for renal outcomes only.17

Given the recent availability of more clinical outcome trials focusing on heart failure and renal outcomes, we conducted a systematic review and performed meta‐analyses and meta‐regression analyses of GLP‐1RA and SLGT2i trials, evaluating the broader scope of relative treatment benefits for atherosclerotic cardiovascular, heart failure, and renal outcomes. Our specific purpose was to increase plausibility of the abovementioned findings from participant‐level mediation analyses for the 5 potential mediators (HbA1c, body weight, systolic blood pressure, hematocrit, and uACR). Using trial‐level data, 2 associations that are required for statistical mediation were assessed. The first corresponds to changes in levels of potential mediators by treatment. The second refers to varying relative treatment benefits with changes in potential mediators.18

Methods

For this study, we expanded the search strategy of a previously published systematic review (International prospective register of systematic reviews [PROSPERO] CRD42022308907) that included 31 reports on 9 GLP‐1RA and 12 SGLT2i trials.19 This prior work included meta‐analyses of hazard ratios extracted from reports of confirmatory randomized controlled trials with atherosclerotic cardiovascular disease (ASCVD), heart failure, and adverse renal outcomes as primary end points. Our expanded search strategy was designed to identify additional articles reporting longitudinal data on the 5 markers. We again followed Preferred Reporting Items for Systematic Reviews and Meta‐Analyses guidelines and registered the study protocol on PROSPERO (CRD42022327225). All supporting data and statistical codes used for our analyses have been made publicly available at GitHub and can be accessed at: https://github.com/Modeling‐NovelDiabetesMeds. This study used aggregate trial data and did not require review by an Institutional Review Board or informed consent.

Search Strategy and Eligibility Criteria

We considered articles eligible for inclusion if trials reported enrolling adult patients with or without diabetes, but not exclusively patients with type 1 or gestational diabetes. Moreover, we required trial designs to randomize patients to an intervention arm in which GLP‐1RAs or SGLT2i were administered as a single drug added to existing therapy and standard of care. The trials' control arm was required to be defined as existing therapy and standard of care with or without placebo. Furthermore, we required reporting of efficacy data for at least 1 of the following outcomes: MACE, myocardial infarction (MI), stroke, all‐cause mortality, cardiovascular mortality, hospitalization for heart failure, or a composite renal outcome. We defined the composite renal outcome as first occurrence of worsening kidney function, end‐stage renal disease, or death from renal causes (Table S1).

To identify additional trial reports of secondary and post hoc analyses that contained information on longitudinal changes of the 5 markers not already reported in the 31 studies of efficacy data, we extended our initial search strategy with more targeted PubMed searches through April 16, 2023. Because most of the primary trial reports already included data on changes in HbA1c, systolic blood pressure, and body weight, these targeted searches were confined to the few trials that did not report these data. In addition, we constructed specific syntax for identifying secondary trial reports on changes in hematocrit and uACR. Example PubMed search syntax for the Dapagliflozin and Prevention of Adverse Outcomes in Chronic Kidney Disease (DAPA‐CKD) trial and Dapagliflozin Evaluation to Improve the Lives of Patients with Preserved Ejection Fraction Heart Failure (DELIVER) trial is detailed in Table S2. Additionally, identified articles were considered eligible if they included longitudinal change or follow‐up score data reported for the entire trial population per study arm. Two independent reviewers (J.M.R.‐V. and B.S.F.) performed screening of title and abstracts, followed by full‐text reviewing for eligibility.

Data Extraction and Quality Assessment

Two reviewers (J.M.R.‐V. and M.T.) independently extracted data from selected full‐text reports. Disagreements were resolved through discussion or arbitration with a third reviewer (B.S.F.). We extracted baseline as well as change or follow‐up score data of potential mediators for each eligible trial per study arm, including 95% CIs, standard errors, or otherwise standard deviations together with the number of participants. We allowed change or follow‐up score data from analyses of crude longitudinal data or repeated measures modeling. We established 1 year as the criterion for selecting the time frame for the change or follow‐up scores. When information was not available in text format, we extracted these data from published plots using an online tool.20 If 1‐year change or follow‐up score information was not available, we chose the immediate available measure before 1 year followed by the next available measure after the 1‐year cutoff. If none of these criteria were present, we extracted the best available data point, either at the end of follow‐up or the mean difference estimated across time using longitudinal modeling.

Statistical Analysis

We followed the theoretical causal structure of mediation analysis described by the directed acyclic graph as shown in Figure 1.21 The indirect treatment effect (path A and B) is the effect of treatment on the outcome derived from treatment‐induced changes in the mediator. The direct treatment effect (path C) is the effect of treatment on the outcome keeping the mediator fixed at the level that corresponds to the control arm. With meta‐analysis we investigated how longitudinal differences in potential mediators vary with treatment (path A). Subsequently, we assessed variation in outcome levels with changes in the potential mediators using meta‐regression analysis (path B). We omitted estimation of direct treatment effects (path C) as well as potential confounding by other time‐varying risk factors because these cannot be addressed appropriately using aggregated trial data only.22

Figure 1. Mediation analysis framework.

GLP‐1RA indicates glucagon‐like peptide‐1 receptor agonist; HbA1c; glycated hemoglobin; MACE, major adverse cardiovascular events; and SGLT2i, sodium‐glucose cotransporter‐2 inhibitor.

For computing longitudinal differences in potential mediators, we subtracted the mean in the control group from the mean in the treatment group. To compute these changes, we chose either the change or follow‐up score depending on data availability. The standard error of the difference was calculated by pooling extracted standard errors from each study arm.23 When data were presented by subgroup, we used fixed‐effect meta‐analysis to obtain information by study arm. Subsequently, we estimated mean differences and their 95% CIs for potential mediators of interest for GLP‐1RA and SGLT2i trials separately using meta‐analysis with a random‐effects model. In this model, the reported effect size of each study was weighted by the inverse of its variance. To evaluate interstudy heterogeneity, we used the Cochran Q and I2 statistics. The Cochran Q test is based on a χ2 distribution, and the null hypothesis assesses a common effect size that all trials share. The I2 statistic quantifies the fraction of the overall observed variance that genuinely reflects disparities in effect size. Typical I2 thresholds are used to classify that the degree of heterogeneity is low (≤25%), moderate (26%–50%), or high (>50%).

To estimate associations of differences in potential mediators with relative treatment effects for adverse cardiovascular and renal events, we performed meta‐regressions using linear mixed‐effects modeling. We used the trials' mean difference in potential mediator as a covariate and the log hazard ratio (HR) as dependent variable in separate models for each drug class. Outcomes for the HRs were categorized as ASCVD events including MACE, fatal or nonfatal MI, fatal or nonfatal stroke, and cardiovascular mortality. Other clinical outcomes were defined as hospitalization for heart failure, the composite renal outcome, and all‐cause mortality.

In addition to estimation of slopes that determine the main association per drug class, we evaluated heterogeneity in slopes by drug class using a separate analysis with combined models that included stratification by drug class and an additional interaction term of drug class and mediator of interest. We tested for differences in slopes (LogHRinteraction) using the likelihood‐ratio test. We required at least 5 studies per drug class in each meta‐regression analysis to assess statistical significance, with P<0.05 as the threshold. To assess the potential impact of differences in trial design and setting, we conducted additional meta‐regression analyses excluding studies that used divergent methods. First, we excluded studies in which changes in potential mediators corresponded to the complete trial duration instead of short‐term. Second, we excluded trials done in the acute care setting, as well as those that exclusively enrolled patients with established heart failure at baseline. Third, we excluded trials with SGLT2i also inhibiting sodium‐glucose cotransporter‐1 (SGLT1). Last, we performed sensitivity analyses excluding trials with an inconsistent end point definition for MACE and the composite renal outcome.

For the analyses, we used the metafor and meta packages in R version 4.1.1 (R Foundation for Statistical Computing; http://www.r‐project.org).

Results

Characteristics of Included Trials

From the total of 318 articles, we ultimately included 51 full‐text reports on 22 trials (n=148 386): 9 on GLP‐1RAs12, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40 (n=64 326) versus 13 on SGLT2i3, 16, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71 (n=84 150) (Figure S1). The mean age of the trial participants ranged between 62 and 72 years, and 36% of the trial participants were women. Baseline HbA1c ranged from 6.6% to 8.9%, and body weight ranged between 88.1 and 91.7 kg. For systolic blood pressure, the baseline values ranged from 122 to 140 mm Hg (Table 1). Hematocrit was assessed longitudinally in 8 SGLT2i trials,16, 41, 42, 44, 47, 48, 49, 50, 52, 54, 55, 56, 57, 58, 59, 71 and its baseline values ranged from 39.1% to 43.3% in these trials. For GLP1‐RAs, hematocrit was generally not reported. For uACR, we performed analyses including 4 GLP‐1RA trials24, 25, 26, 27, 28, 29, 36, 37 that reported longitudinal data allowing for estimation of the absolute difference in geometric means. These GLP‐1RA trials included heterogeneous patient populations with and without albuminuric chronic kidney disease. However, for SGLT2i trials, data on a consistent measure for longitudinal uACR were generally not available (Table S3). Among 4 SGLT2i trials reporting longitudinal geometric mean values, 2 of the trials50, 52 enrolled a large number of patients with albuminuric chronic kidney disease, and therefore, changes were not comparable. Baseline uACR ranged from 15.9 to 28.3 mg/g in the 4 analyzed GLP‐1RA trials (Table 1), which included 25% to 100% with albuminuria at baseline (Table S4). The range of baseline uACR overlapped with the other 5 GLP‐1RA trials. All included trials were deemed high quality and with a low risk of bias (Tables S5 and S6).

TABLE 1 Characteristics of GLP‐1RA and SGLT2i Trials

TrialsDrug, dose (route)Trial duration, yAge, yWomen, N (%)Diabetes, N (%)Diabetes duration, yASCVD/HF, N (%)HbA1c, %Body weight, kgSBP, mm HgHematocrit, %uACR, mg/g
GLP‐1RA trials
ELIXA24, 25 (n=6068)Lixisenatide, 10 μg/d or 20 μg/d (SQ)2.160 (10)1861 (31)6068 (100)9.2 (8.2)6068 (100)/1358 (22)7.7 (1.3)84.9 (19.4)129.5 (17.0)10.4 (6.0–31.4)18.1
LEADER26, 27, 28 (n=9340)Liraglutide, 1–8 mg/d (SQ)3.864 (7)3337 (36)9340 (100)12.8 (8.0)7598 (81)/1667 (18)8.7 (1.6)91.7 (21.0)135.9 (17.8)NA20.4
SUSTAIN‐628, 29, 30 (n=3297)Semaglutide, 0–5 mg/wk or 1 mg/wk (SQ)2.165 (7)1295 (39)3297 (100)13.9 (8.1)2735 (83)/777 (24)8.7 (1.5)92.1 (20.6)135.6 (17.2)NA23.4
EXSCEL31, 32 (n=14 752)Exenatide, 2 mg/wk (SQ)3.262 (9)5603 (38)14 752 (100)13.1 (8.3)10 782 (73)/2389 (16)8.192.3135.5NANA
FREEDOM‐CVO33 (n=4156)Exenatide, continuous osmotic minipump (SQ)1.363 (57–68)1525 (37)4156 (100)10.3 (5.9–15.4)3159 (76)/668 (16)8.0NR137.5NANA
Harmony Outcomes34, 35 (n=9463)Albiglutide, 30 mg/wk or 50 mg/wk (SQ)1.564 (7)2894 (31)9463 (100)14.2 (8.8)9463 (100)/1922 (20)8.7 (1.5)92.2 (20.1)134.8 (16.6)NANA
REWIND12, 36, 37 (n=9901)Dulaglutide, 1–5 mg/wk (SQ)5.466 (7)4589 (46)9901 (100)10.5 (7.2)3114 (31)/853 (9)7.4 (1.1)88.6137.2 (16.8)NA15.9 (6.2–61)
PIONEER 638 (n=3183)Semaglutide, 14 mg/d (oral)1.366 (7)1007 (32)3183 (100)14.9 (8.5)2695 (85)/388 (12)8.2 (1.6)90.9 (21.2)135.5 (18.0)NANA
AMPLITUDE‐O39, 40 (n=4076)Efpeglenatide, 4 mg/wk or 6 mg/wk (SQ)1.865 (8)1344 (33)4076 (100)15.4 (8.8)3650 (90)/737 (18)8.9 (1.5)NR134.9 (15.5)NA28.3 (9.7–114.2)
SGLT2i trials
EMPA‐REG Outcome16, 41, 42, 43 (n=7020)Empagliflozin, 10 mg/d or 25 mg/d (oral)3.163 (9)2004 (29)7020 (100)57%>10†7020 (100)/706 (10)8.0 (0.8)86.3 (19.0)135.4 (17.0)43.3NA
CANVAS Program44, 45, 46 (n=10 142)Canagliflozin, 100 mg/d (oral)2.463 (8)3632 (36)10 142 (100)13.5 (7.8)6656 (66)/1461 (14)NR90.2136.6 (15.8)NA12.3
DECLARE‐TIMI 5847, 48, 49 (n=17 160)Dapagliflozin, 10 mg/d (oral)4.264 (7)6422 (37)17 160 (100)11.8 (7.8)6974 (41)/1724 (10)8.3 (1.2)90.7135.0 (15.4)NANA
CREDENCE50, 51, 52, 53 (n=4401)Canagliflozin, 100 mg/d (oral)2.663 (9)1494 (34)4401 (100)15.8 (8.6)2220 (50)/652 (15)8.3 (1.3)NR140.0 (15.6)NA927
VERTIS CV54, 55 (n=8246)Ertugliflozin, 5 mg/d or 15 mg/d (oral)3.064 (8)2477 (30)8246 (100)13.0 (8.3)8246 (100)/1958 (24)8.2 (1.0)NR133.4 (13.8)NANA
DAPA‐HF56, 57, 58 (n=4744)Dapagliflozin, 10 mg (oral)1.367 (11)1109 (23)1993 (42)7.4 (2.73–13.5)NR/4744 (100)NRNR122.041.1NA
EMPEROR‐Preserved59, 60 (n=5988)Empagliflozin, 10 mg (oral)2.272 (9)2676 (45)2934 (49)NRNR/5987 (100)7.3 (1.5)82 (19)131.9 (16)NRNA
DAPA‐CKD61, 62, 63 (n=4304)Dapagliflozin, 10 mg (oral)2.462 (12)1425 (33)2888 (67)13.7 (7.1–20.0)1610 (37)/468 (11)7.0 (1.4)81.8 (20.5)137.1 (17.4)NA950 (472–1903)
EMPEROR‐Reduced60, 64, 65 (n=3730)Empagliflozin, 10 mg (oral)1.367 (11)893 (24)1865 (50)NRNR/3730 (100)6.6 (1.2)NR122.0 (15.6)39.6NA
SOLOIST‐WHF66 (n=1222)Sotagliflozin, 200–400 mg (oral)0.869 (63–76)412 (34)1222 (100)10.2 (5.0–16.9)NR/1222 (100)7.2NR122NANA
SCORED67 (n=10 584)Sotagliflozin, 200–400 mg (oral)1.369 (63–74)4754 (45)10 584 (100)NR5429 (51)/3283 (31)8.3 (7.6–9.4)NR139 (127–149)NA75 (17–486)
DELIVER3, 68, 69, 70 (n=6263)Dapagliflozin, 10 mg (oral)2.372 (10)2747 (44)2806 (45)NR6263 (100)/6263 (100)NANR128 (15)NANA
EMPA‐KIDNEY71 (n=6609)Empagliflozin, 10 mg (oral)2.064 (10)2192 (33)3040 (46)NR1765 (26)/658 (9)NANA136.6 (18.3)39.1 (5.1)329 (46–1074)

Data are mean (SD), N (%), or median (interquartile range), unless otherwise specified.

AMPLITUDE‐O indicates effect of efpeglenatide on cardiovascular outcomes; ASCVD, atherosclerotic cardiovascular disease; CANVAS, canagliflozin cardiovascular assessment study; DAPA‐HF, dapagliflozin and prevention of adverse outcomes in heart failure; CREDENCE, evaluation of the effects of canagliflozin on renal and cardiovascular outcomes in participants with diabetic nephropathy; DAPA‐CKD, a study to evaluate the effect of dapagliflozin on renal outcomes and cardiovascular mortality in patients with chronic kidney disease; DECLARE‐TIMI 58, multicenter trial to evaluate the effect of dapagliflozin on the incidence of cardiovascular events; DELIVER, dapagliflozin evaluation to improve the lives of patients with preserved ejection fraction heart failure; DPP‐4, dipeptidyl peptidase 4; ELIXA, evaluation of lixisenatide in acute coronary syndrome; EMPA‐KIDNEY, the study of heart and kidney protection with empagliflozin; EMPA‐REG Outcome, BI 10773 (empagliflozin) cardiovascular outcome event trial in type 2 diabetes mellitus patients; EMPEROR‐Preserved, empagliflozin outcome trial in patients with chronic heart failure with preserved ejection fraction; EMPEROR‐Reduced, empagliflozin outcome trial in patients with chronic heart failure and a reduced ejection fraction; EXSCEL, exenatide study of cardiovascular event lowering; FREEDOM‐CVO, study to evaluate cardiovascular outcomes with ITCA 650 in patients treated with standard of care for type 2 diabetes; GLP‐1RA, glucagon‐like peptide‐1 receptor agonist; Harmony Outcomes, effect of albiglutide, when added to standard blood glucose lowering therapies, on major cardiovascular events in subjects with type 2 diabetes mellitus; HbA1c, glycated hemoglobin; HF, heart failure; LEADER, liraglutide effect and action in diabetes: Evaluation of cardiovascular outcome results; NA, not applicable; NR, not reported; PIONEER 6, peptide innovation for early diabetes treatment 6; REWIND, researching cardiovascular events with a weekly incretin in diabetes; SBP, systolic blood pressure; SCORED, effect of sotagliflozin on cardiovascular and renal events in participants with type 2 diabetes and moderate renal impairment who are at cardiovascular risk; SGLT2i, sodium‐glucose cotransporter‐2 inhibitor; SOLOIST‐WHF, effect of sotagliflozin on cardiovascular events in participants with type 2 diabetes post worsening heart failure; SQ, subcutaneous; SUSTAIN‐6, trial to evaluate cardiovascular and other long‐term outcomes with semaglutide in subjects with type 2 diabetes; uACR, urine albumin‐creatinine ratio and VERTIS CV, evaluation of ertugliflozin efficacy and safety cardiovascular outcomes trial.

The majority of longitudinal difference measures (N=51 out of 110 trials risk factors) could be obtained at 50 to 52 weeks of follow‐up. For N=10 out of 110, the time point for estimation was <50 weeks. For N=16 out of 110, estimates were obtained as time‐weighted average over the entire follow‐up period by repeated measure analyses (Table S7).

Meta‐Analysis of Mean Differences in Potential Mediators

For GLP‐1RA trials, the summary mean difference in HbA1c was −0.75% (95% CI, −1.06 to −0.45), for body weight −2.08 kg (95% CI, −3.13 to −1.03), for systolic blood pressure −1.46 mm Hg (95% CI, −2.88 to −0.04), and for uACR it was −4.88 mg/g (95% CI, −9.60 to 0.17). For SGLT2i trials, the mean difference in HbA1c was −0.33% (95% CI, −0.48 to −0.20), for body weight −1.47 kg (95% CI, −1.94 to −0.99), and for systolic blood pressure −2.41 mm Hg (95% CI, −3.20 to −1.63). Hematocrit increased with SGLT2i by 2.53% (95% CI, 2.36–2.70) on average (see Figure 2). Among trials for both drug classes, the degree of heterogeneity in these mean difference estimates was high, with all I2 statistics >50% and Q statistics P<0.01.

Figure 2. Forest plots of mean differences in potential mediators among GLP‐1RA and SGLT2i trials.

AMPLITUDE‐O indicates effect of efpeglenatide on cardiovascular outcomes; CANVAS, canagliflozin cardiovascular assessment study; CREDENCE, evaluation of the effects of canagliflozin on renal and cardiovascular outcomes in participants with diabetic nephropathy; DAPA‐HF, dapagliflozin and prevention of adverse outcomes in heart failure; DECLARE‐TIMI 58, multicenter trial to evaluate the effect of dapagliflozin on the incidence of cardiovascular events; DELIVER, dapagliflozin evaluation to improve the lives of patients with preserved ejection fraction heart failure; DPP‐4, dipeptidyl peptidase 4; EMPA‐KIDNEY, the study of heart and kidney protection with empagliflozin; EMPA‐REG Outcome, BI 10773 (empagliflozin) cardiovascular outcome event trial in type 2 diabetes mellitus patients; EMPEROR‐Preserved, empagliflozin outcome trial in patients with chronic heart failure with preserved ejection fraction; EMPEROR‐Reduced, empagliflozin outcome trial in patients with chronic heart failure and a reduced ejection fraction; ELIXA, evaluation of lixisenatide in acute coronary syndrome; FREEDOM‐CVO, study to evaluate cardiovascular outcomes with itca 650 in patients treated with standard of care for type 2 diabetes; GLP‐1RA glucagon‐like peptide‐1 receptor agonist; Harmony Outcomes, effect of albiglutide, when added to standard blood glucose lowering therapies, on major cardiovascular events in subjects with type 2 diabetes mellitus; HK, Hartung‐Knapp adjustment; LEADER, liraglutide effect and action in diabetes: evaluation of cardiovascular outcome results; MD, mean difference; PIONEER 6, peptide innovation for early diabetes treatment 6; REWIND, researching cardiovascular events with a weekly incretin in diabetes; SCORED, effect of sotagliflozin on cardiovascular and renal events in participants with type 2 diabetes and moderate renal impairment who are at cardiovascular risk; SGLT2i, sodium‐glucose cotransporter‐2 inhibitor; SUSTAIN‐6, trial to evaluate cardiovascular and other long‐term outcomes with semaglutide in subjects with type 2 diabetes; uACR, urine albumin‐creatinine ratio and VERTIS CV, evaluation of ertugliflozin efficacy and safety cardiovascular outcomes trial.

Meta‐Regression of Differences in Potential Mediators and Relative Treatment Effects

HRs for MACE, MI, stroke, and cardiovascular mortality improved with more HbA1c reduction among GLP‐1RA trials (ΔHR, 21%–30%; ΔLogHR, 0.23–0.36). The association for MACE (ΔHR, 23%; ΔLogHR, 0.26) was statistically significant (P=0.02). Among SGLT2i trials, associations between HbA1c reduction and ASCVD outcomes were generally in the opposite direction, and HRs became less protective with more HbA1c reduction (ΔHR, 0% to −97%; ΔLogHR, 0 to −0.68), although these trends were not statistically significant. Associations were significantly different between drug classes for MACE (Pinteraction=0.04; Table 2).

Table 2. Meta‐Regression Coefficients for Log HRs by Mean Differences in Potential Mediators Among GLP‐1RA and SGLT2i Trials

MACEMyocardial infarctionStrokeCardiovascular mortalityHospitalization for heart failureComposite renal outcomeAll‐cause mortality
HbA1c
∆LogHRGLP1‐RA0.26* (0.05 to 0.48)0.23 (−0.08 to 0.54)0.36 (−0.10 to 0.82)0.23 (−0.12 to 0.58)0.14 (−0.25 to 0.54)0.14 (−0.29 to 0.57)0.07 (−0.21 to 0.36)
PGLP1‐RAs0.020.140.130.190.480.530.62
∆LogHRSGLT2i−0.20 (−0.62 to 0.22)−0.31 (−1.21 to 0.59)−0.68 (−1.94 to 0.57)0.00 (−0.45 to 0.44)−0.10 (−0.48 to 0.28)0.22 (−0.43 to 0.87)−0.03 (−0.46 to 0.40)
PSGLT2i0.360.510.290.990.590.510.88
∆LogHRinteraction−0.47* (−0.92 to −0.01)−0.54 (−1.47 to 0.40)−0.99 (−2.05 to 0.06)−0.24 (−0.76 to 0.28)−0.24 (−0.79 to 0.30)0.06 (−0.72 to 0.84)−0.09 (−0.58 to 0.40)
Pinteraction0.040.220.070.330.380.790.70
Body weight
∆LogHRGLP1‐RA0.06 (−0.02 to 0.13)0.04 (−0.06 to 0.14)0.07 (−0.05 to 0.19)0.07 (−0.03 to 0.17)−0.01 (−0.11 to 0.09)0.00 (−0.12 to 0.11)0.02 (−0.06 to 0.10)
PGLP1‐RA0.160.420.230.150.870.960.58
∆LogHRSGLT2i−0.14 (−0.28 to 0.01)−0.25* (−0.46 to −0.05)−0.29 (−0.59 to 0.01)0.01 (−0.13 to 0.15)0.03 (−0.10 to 0.15)0.10 (−0.10 to 0.30)0.03 (−0.10 to 0.15)
PSGLT20.060.020.060.870.660.320.66
∆LogHRinteraction−0.20* (−0.37 to −0.02)−0.30* (−0.53 to −0.08)−0.35* (−0.64 to −0.06)−0.07 (−0.22 to 0.09)0.04 (−0.12 to 0.20)0.10 (−0.13 to 0.33)0.00 (−0.13 to 0.14)
Pinteraction0.020.010.020.360.660.211.00
Systolic blood pressure
∆LogHRGLP1‐RA−0.02 (−0.11 to 0.06)−0.08 (−0.17 to 0.01)0.03 (−0.09 to 0.15)0.04 (−0.07 to 0.14)−0.12 (−0.23 to 0.00)−0.09 (−0.30 to 0.11)0.04 (−0.05 to 0.12)
PGLP1‐RA0.600.090.620.480.060.360.37
∆LogHRSGLT2i0.01 (−0.10 to 0.11)−0.03 (−0.19 to 0.13)−0.02 (−0.28 to 0.23)0.02 (−0.04 to 0.08)0.02 (−0.03 to 0.08)0.09 (0.00 to 0.19)0.03 (−0.02 to 0.08)
PSGLT2i0.920.710.850.490.470.060.22
∆LogHRinteraction0.03 (−0.11 to 0.17)0.05 (−0.12 to 0.22)−0.05 (−0.28 to 0.18)−0.02 (−0.14 to 0.11)0.14* (0.01 to 0.27)0.18 (−0.06 to 0.43)−0.01 (−0.10 to 0.09)
Pinteraction0.410.440.770.750.040.090.88
Hematocrit
∆LogHRSGLT2i0.13 (−0.44 to 0.71)0.31 (−0.40 to 1.03)0.80 (−0.12 to 1.71)−0.21 (−0.67 to 0.25)−0.08 (−0.41 to 0.25)−0.27 (−0.91 to 0.38)−0.23 (−0.58 to 0.12)
PSGLT20.650.390.090.370.630.420.21
uACR
∆LogHRGLP1‐RA0.03 (−0.01 to 0.08)0.04 (−0.01 to 0.09)0.04 (−0.05 to 0.13)0.02 (−0.04 to 0.08)−0.01 (−0.06 to 0.05)−0.01 (−0.09 to 0.06)0.00 (−0.04 to 0.05)
PGLP1‐RANANANANANANANA

Meta‐regression coefficients were computed using the log HR as dependent variable. ∆LogHRGLP1‐RA and ∆LogHRSGLT2i are model coefficients from analyses stratified by GLP‐1RA vs. SGLT2i trials. ∆LogHRinteraction are model coefficients from analyses that additionally included an interaction term of drug class strata and mediator of interest. When these slope estimates are exponentiated, they represent the percent change in the geometric average hazard ratio that can be expected when the difference in potential mediator follow‐up levels between study arms is changed by 1 unit. GLP‐1RA indicates glucagon‐like peptide‐1 receptor agonist; HR, hazard ratio; MACE, major adverse cardiovascular events; SGLT2i, sodium‐glucose contraposer‐2 inhibitor; and uACR, urine albumin‐creatinine ratio.

*P<0.05.

Slopes, but not P‐values are reported for meta‐regressions were the number of trials with available data was less than five.

Hazard ratios for MACE, MI, and stroke became more protective with more weight loss in GLP‐1RA trials (ΔHR, 4%–7%; ΔLogHR, 0.04–0.07; P>0.05) but not for SGLT2i (ΔHR, −15%, −28%, and −34%; ΔLogHR, −0.14, −0.25, and −0.29; P=0.2, 0.6, and 0.2, respectively). Meta‐regression associations differed significantly between drug classes (Pinteraction<0.05). For stroke, HRs tended to become less protective, with larger increases in hematocrit among the 5 SGLT2i trials with available data (ΔHR, 123%; ΔLogHR, 0.80; P=0.09). We did not find any indication of association with systolic blood pressure and uACR for GLP‐1RAs or SGLT2i for any outcome. Changes in HRs varied from −9% to 11% (ΔLogHRs, −0.12 to 0.09) and none were statistically significant (Figures 3, 4, 5, 6 through 7; Table 2). Associations for hospitalization with heart failure, composite renal outcome, and all‐cause mortality were not statistically significant for any of the potential mediators (Figures 6 and 7, Table 2).

Figure 3. Treatment effects for MACE by mean difference in potential mediators.

The size of each trial's circle is proportional to the inverse of the variance of its log hazard ratio. The line and shaded area refer to point estimates and 95% confidence bands from the meta‐regression analyses. A solid line refers to meta‐regression slopes with statistical significance. For uACR the figure represents an association, but no statistical testing was performed due to the small number of included GLP1‐RA trials. GLP‐1RA indicates glucagon‐like peptide‐1 receptor agonist; HbA1c, glycated hemoglobin; MACE, major adverse cardiovascular events; SGLT2i, sodium‐glucose cotransporter‐2 inhibitor; and uACR, urine albumin‐creatinine ratio.

Figure 4. Treatment effects for myocardial infarction and stroke by mean difference in potential mediators.

The size of each trial's circle is proportional to the inverse of the variance of its log hazard ratio. The line and shaded area refer to point estimates and 95% confidence bands from the meta‐regression analyses. A solid line refers to meta‐regression slopes with statistical significance. For uACR, the figure represents an association, but no statistical testing was performed due to the small number of included GLP1‐RA trials. GLP‐1RA indicates glucagon‐like peptide‐1 receptor agonist; HbA1c, glycated hemoglobin; SGLT2i, sodium‐glucose cotransporter‐2 inhibitor; and uACR, urine albumin‐creatinine ratio.

Figure 5. Treatment effects for cardiovascular and all‐cause mortality by mean difference in potential mediators.

The size of each trial's circle is proportional to the inverse of the variance of its log hazard ratio. The line and shaded area refer to point estimates and 95% confidence bands from the meta‐regression analyses. For uACR the figure represents an association, but no statistical testing was performed due to the small number of included GLP1‐RA trials. GLP‐1RA indicates glucagon‐like peptide‐1 receptor agonist; HbA1c, glycated hemoglobin; SGLT2i, sodium‐glucose cotransporter‐2 inhibitor; and uACR, urine albumin‐creatinine ratio.

Figure 6. Treatment effects for hospitalization for heart failure by mean difference in potential mediators.

The size of each trial's circle is proportional to the inverse of the variance of its log hazard ratio. The line and shaded area refer to point estimates and 95% confidence bands from the meta‐regression analyses. For uACR the figure represents an association, but no statistical testing was performed due to the small number of included GLP1‐RA trials. GLP‐1RA indicates glucagon‐like peptide‐1 receptor agonist; HbA1c, glycated hemoglobin; SGLT2i, sodium‐glucose cotransporter‐2 inhibitor; and uACR, urine albumin‐creatinine ratio.

Figure 7. Treatment effects for the composite renal outcome by mean difference in potential mediators.

The size of each trial's circle is proportional to the inverse of the variance of its log hazard ratio. The line and shaded area refer to point estimates and 95% confidence bands from the meta‐regression analyses. For uACR the figure represents an association, but no statistical testing was performed due to the small number of included GLP1‐RA trials. GLP‐1RA indicates glucagon‐like peptide‐1 receptor agonist; HbA1c, glycated hemoglobin; SGLT2i, sodium‐glucose cotransporter‐2 inhibitor; and uACR, urine albumin‐creatinine ratio.

After excluding trials with divergent criteria for potential mediator differences, trial populations, and renal end point definitions, the magnitude and direction of slopes remained generally unchanged (Tables S8–S14).

Discussion

In this study, we summarized data from 22 trials (n=148 386): 9 on GLP‐1RAs (n=64 326) and 13 on SGLT2i (n=84 150). We found significant, but heterogeneous effects on longitudinal measurements of all 5 potential mediators (HbA1c, body weight, systolic blood pressure, hematocrit, and uACR) among trials evaluating both drug classes. In meta‐regression analyses of 1‐year mean differences in potential mediators versus log HRs, in contrast to findings for SGLT2i trials, treatment effects for MACE became more protective, with more HbA1c reduction among GLP‐1RA trials. For body weight, treatment effects for MACE, MI, and stroke showed diminishing trends with more weight loss for SGLT2i but not for GLP‐1RA trials. The relationship between a larger decrease in body weight and reduced treatment efficacy for MI was statistically significant for SGLT2i. Among the 5 SGLT2i trials with available data on both hematocrit and any stroke rates, HRs trended >1 with larger increases in hematocrit, indicating a potential harmful effect of increasing hematocrit. However, these associations were not statistically significant. For heart failure and renal outcomes, no significant associations were found. No indication for potential mediation by changes in systolic blood pressure and uACR were found for any of the clinical outcomes.

Our finding of reductions in HbA1c as a potential mediator for effects on MACE by GLP‐1RAs but not SGLT2i corroborates with the results from published participant‐level mediation analyses.10, 11, 12, 16 In addition, meta‐regression analyses of various glycemic control trials have shown a significant correlation between lowering HbA1c and reduced MACE rates.72 The summary estimate for mean HbA1c reduction appeared on average larger among GLP‐1RA trials as compared with SGLT2i trials. However, it should be acknowledged that change in HbA1c as mediator of benefits for ASCVD events has been considered controversial based on individually published findings from randomized controlled trials that evaluated intensive glycemic control.73, 74, 75 These findings were not supportive of improved efficacy with more HbA1c reduction across all patients. Explanations for differences in efficacy include heterogeneity of trial populations for history of glycemic control and hypoglycemic events, degree of established ASCVD, and aging and frailty.76 The 2 trials that showed the largest reduction in HbA1c reduction in our analysis were GLP‐1RA trials (Effect of Efpeglenatide on Cardiovascular Outcomes [AMPLITUDE‐O] and Trial to Evaluate Cardiovascular and Other Long‐term Outcomes With Semaglutide in Subjects With Type 2 Diabetes [SUSTAIN‐6]). These trials also exhibited the largest relative reduction in MACE rates with HRs of 0.73 (95% CI, 0.58–0.92) and 0.74 (95% CI, 0.58–0.95), but they also had the highest mean baseline HbA1c levels of 8.9% and 8.7%, respectively. It could be hypothesized that the larger HbA1c reductions in these trials were accompanied with lower risk of hypoglycemia than observed in the other trials, given the higher baseline values. Nevertheless, it remains important to note that the possibility that HbA1c rather is a marker of other underlying mediating factors also affected by GLP1‐RAs cannot be precluded.10

However, in contrast to our findings, other meta‐regression studies of aggregated trial data showed that weight reduction was not correlated with diminished MACE rate reductions but demonstrated it was an important marker of greater reductions in heart failure event rates.77, 78, 79, 80 Several reasons can be mentioned for the conflicting results. First, these prior meta‐regression analyses combined data from GLP‐1RA and SGLT2i trials, as well as trials of other intervention types without adjustment or stratification. Second, sustained weight reduction induced by GLP‐1RAs and SGLT2i predominantly related to caloric loss and reduction of visceral fat and less to changes in fluid status, weakening the evidence for weight loss as a protector against heart failure.8 On the other hand, caloric loss with these agents may result in excessive low body weight or compensatory food intake in some patients, both of which could explain a potential association with increased MACE rates.9, 81

Also, the trend we observed for increases in hematocrit and HRs for stroke conflicted with the evidence for hematocrit as a marker of improved nutrient deprivation signaling with SGLT2i. Some have argued that high hematocrit might be reflective of prolonged greater volume depletion and erythrocytosis, which may lead to increased blood viscosity and reduced perfusion of ischemic tissues.82, 83, 84 Findings of higher amputation risk and lower efficacy for MACE events that are ischemic of nature may be supportive of this hypothesis.85, 86 However, just like for changes in HbA1c, it is difficult to assess the underlying mechanism for these findings. For example, heterogeneity in study populations, differences in trial design, drug‐specific differences of mediation among the same drug class, and changes in other risk factor levels induced by treatment may be explanatory.

Compared with the prior meta‐regression analyses investigating mediation mechanisms for GLP‐1RAs and SGLT2i, we were able to include data from newer, more recently published trial reports providing up‐to‐date and more precise data. As such, the main strength of our study is that we could evaluate heterogeneity in effects by changes in potential mediators for multiple clinical outcomes and testing for effect modification by drug class. In addition, we assessed longitudinal differences in potential mediators across study arms at a predefined time point of 1 year. A 1‐year interval has also been used by others to define baseline periods for investigating the impact of changes in markers of cardiovascular and renal outcomes within a meta‐analytic framework. For example, 1‐year changes in albuminuria were evaluated as markers of subsequent risk of end‐stage kidney disease in 2 large individual participant‐level meta‐analyses.87, 88 In addition, mean absolute changes in low‐density lipoprotein cholesterol within statin trials were linked to relative reductions in vascular events in meta‐analyses conducted by the Cholesterol Treatment Trialists' Collaboration.89 Using changes over the entire trial duration may lead to biases due to reverse causality and survivor selection.

The strengths of our study should be viewed in the context of some limitations. First, we could not always estimate 1‐year differences in mean achieved levels of potential mediators between the treatment and control groups for all trials. For example, due to differences in study design, mean differences in hematocrit were only available for SGLT2i trials, whereas the summary estimate for uACR could only be estimated using a limited number of GLP‐1RA trials. However, we generally included at least 5 trials when performing the analysis, a reasonable sample size for meta‐analysis and meta‐regression analysis.90 In addition, our findings were robust against exclusion of trials that reported changes over a shorter or longer period of time. Second, we used combined end point definitions of MACE, stroke, and renal outcomes that may reflect a weighted average of different associations for subevent types. For example, we used a combined stroke end point, whereas for SGLT2i, HRs may be only significant for hemorrhagic strokes.86 If mediators act only on 1 or a few of the single end points within the combined end point, these associations may be missed. Third, we selected 5 potential mediators based on statistical evidence for mediation. However, such statistical evidence does not necessarily imply biological mediation. For example, changes in uACR likely reflect concomitant improvements in endothelial and glycocalyx function, which may have explained the discovery of significant mediation proportions in participant‐level mediation analyses for GLP1‐RAs and SGLT2i.10, 12, 13, 14, 15, 16 Fourth, observed ecological associations may have originated from interstudy heterogeneity due to differences in other risk factors or mediators that were not considered in our analyses. We preferred to focus our analyses on 5 potential mediators that were already identified by participant‐level analyses, presumably limiting the chance of biased findings as well as false discovery. As such, we deemed it relevant to also describe trends that were statistically nonsignificant in our analysis, especially considering the included number of studies. Finally, we limited our search strategy to PubMed as defined by our prior work,19 and as such we may have missed data in reports of secondary analyses for some of the included clinical outcome trials.

We attempted to increase credibility of evidence on the potential mediating role of 5 markers in explaining the clinical effects of 2 novel diabetes agents, GLP‐1RAs and SGLT2is. For this purpose, we analyzed aggregated trial data assessing 2 important components of the indirect treatment effect that is used to identify mediation. It is theoretically possible to perform a meta‐analysis of trial‐level indirect treatment effects using reasonably uniform methods.22 However, findings from the published mediation analyses were generally limited by variation in statistical methodologies and trial design. Initiatives such as the SGLT2i Meta‐Analysis Cardio‐Renal Trialists' Consortium3 that use a standardized protocol and individual participant data collection could be valuable for obtaining mediation estimands.

Mediation meta‐analyses using individual participant data may help verify findings from smaller mechanistic experimental studies to solve remaining uncertainties about the working mechanisms of these agents that drive cardiovascular and renal outcome benefits. While awaiting such analyses, our research suggests potentially important clinical implications, for example, the prospect of achieving more significant reductions in MACE when HbA1c levels show greater improvement with GLP1‐RAs. Additionally, caution may be advised when prescribing SGLT2i medications leads to increased weight loss and hematocrit levels in patients at elevated risk of atherothrombotic events such as ischemic stroke and peripheral artery disease.

We confirm previous findings of increased effectiveness for MACE with reductions in HbA1c for GLP1‐RAs. Further research is needed to investigate the potential loss of SGLT2i efficacy for atherosclerotic cardiovascular outcomes with greater weight loss and increase in hematocrit.

Sources of Funding

This research was supported by the National Heart, Lung, and Blood Institute of the National Institutes of Health under award number R01HL153456. M.G.M.H. has received research funding from the Netherlands Organization for Health Research and Development, the German Innovation Fund, and the Netherlands Educational Grant (Studie Voorschot Middelen), and additional research funding from the Gordon and Betty Moore Foundation. K.E.F. is the recipient of an Innovative Clinical or Translational Science award from the American Diabetes Association (1‐18‐ICTS‐041). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Disclosures

M.G.M.H. receives royalties from Cambridge University Press for a textbook on medical decision making, reimbursement of expenses from the European Society of Radiology for work on the European Society of Radiology guidelines for imaging referrals, and reimbursement of expenses from the European Institute for Biomedical Imaging Research for membership on the Scientific Advisory Board. The remaining authors have no disclosures to report.

Footnotes

* Correspondence to: José M. Rodriguez‐Valadez, PhD, and Bart S. Ferket, MD, PhD, Institute for Healthcare Delivery Science, Department of Population Health Science and Policy, Icahn School of Medicine at Mount Sinai. One Gustave L. Levy Place, Box 1077, New York, NY 10029‐6574. Email: ,

Supplemental Material is available at https://www.ahajournals.org/doi/suppl/10.1161/JAHA.123.032463

This article was sent to Meng Lee, MD, Guest Editor, for review by expert referees, editorial decision, and final disposition.

For Sources of Funding and Disclosures, see page 13.

See Editorial by Dakroub and Almarzooq.

References

  • 1 McGuire DK, Shih WJ, Cosentino F, Charbonnel B, Cherney DZI, Dagogo‐Jack S, Pratley R, Greenberg M, Wang S, Huyck S, et al. Association of SGLT2 inhibitors with cardiovascular and kidney outcomes in patients with type 2 diabetes: a meta‐analysis. JAMA Cardiol. 2021; 6:148–158. doi: 10.1001/jamacardio.2020.4511CrossrefMedlineGoogle Scholar
  • 2 Sattar N, Lee MMY, Kristensen SL, Branch KRH, Del Prato S, Khurmi NS, Lam CSP, Lopes RD, McMurray JJV, Pratley RE, et al. Cardiovascular, mortality, and kidney outcomes with GLP‐1 receptor agonists in patients with type 2 diabetes: a systematic review and meta‐analysis of randomised trials. Lancet Diabetes Endocrinol. 2021; 9:653–662. doi: 10.1016/S2213-8587(21)00203-5CrossrefMedlineGoogle Scholar
  • 3 Nuffield Department of Population Health Renal Studies Group ; SGLT2 inhibitor Meta‐Analysis Cardio‐Renal Trialists' Consortium. Impact of diabetes on the effects of sodium glucose co‐transporter‐2 inhibitors on kidney outcomes: collaborative meta‐analysis of large placebo‐controlled trials. Lancet. 2022; 400:1788–1801. doi: 10.1016/S0140-6736(22)02074-8CrossrefMedlineGoogle Scholar
  • 4 American Diabetes Association Professional Practice Committee ; Draznin B, Aroda VR, Bakris G, Benson G, Brown FM, Freeman R, Green J, Huang E, et al. 10. Cardiovascular disease and risk management: standards of medical care in diabetes‐2022. Diabetes Care. 2022; 45:S144–S174. doi: 10.2337/dc22-S010CrossrefMedlineGoogle Scholar
  • 5 Davies MJ, Aroda VR, Collins BS, Gabbay RA, Green J, Maruthur NM, Rosas SE, Del Prato S, Mathieu C, Mingrone G, et al. A consensus report by the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetes Care. 2022;45:2753–2786. doi: 10.2337/dci22-0034CrossrefMedlineGoogle Scholar
  • 6 Wu Q, Li D, Huang C, Zhang G, Wang Z, Liu J, Yu H, Song B, Zhang N, Li B, et al. Glucose control independent mechanisms involved in the cardiovascular benefits of glucagon‐like peptide‐1 receptor agonists. Biomed Pharmacother. 2022; 153:113517. doi: 10.1016/j.biopha.2022.113517CrossrefMedlineGoogle Scholar
  • 7 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. doi: 10.1016/j.ijcha.2021.100725CrossrefMedlineGoogle Scholar
  • 8 Packer M. Critical reanalysis of the mechanisms underlying the cardiorenal benefits of SGLT2 inhibitors and reaffirmation of the nutrient deprivation signaling/autophagy hypothesis. Circulation. 2022; 146:1383–1405. doi: 10.1161/CIRCULATIONAHA.122.061732LinkGoogle Scholar
  • 9 Avogaro A, de Kreutzenberg SV, Morieri ML, Fadini GP, Del Prato S. Glucose‐lowering drugs with cardiovascular benefits as modifiers of critical elements of the human life history. Lancet Diabetes Endocrinol. 2022; 10:882–889. doi: 10.1016/S2213-8587(22)00247-9CrossrefMedlineGoogle Scholar
  • 10 Buse JB, Bain SC, Mann JFE, Nauck MA, Nissen SE, Pocock S, Poulter NR, Pratley RE, Linder M, Monk Fries T, et al. Cardiovascular risk reduction with liraglutide: an exploratory mediation analysis of the LEADER trial. Diabetes Care. 2020; 43:1546–1552. doi: 10.2337/dc19-2251CrossrefMedlineGoogle Scholar
  • 11 Mann JFE, Buse JB, Idorn T, Leiter LA, Pratley RE, Rasmussen S, Vilsboll T, Wolthers B, Perkovic V. Potential kidney protection with liraglutide and semaglutide: exploratory mediation analysis. Diabetes Obes Metab. 2021; 23:2058–2066. doi: 10.1111/dom.14443CrossrefMedlineGoogle Scholar
  • 12 Konig M, Riddle MC, Colhoun HM, Branch KR, Atisso CM, Lakshmanan MC, Mody R, Raha S, Gerstein HC. Exploring potential mediators of the cardiovascular benefit of dulaglutide in type 2 diabetes patients in REWIND. Cardiovasc Diabetol. 2021; 20:194. doi: 10.1186/s12933-021-01386-4CrossrefMedlineGoogle Scholar
  • 13 Li J, Neal B, Perkovic V, de Zeeuw D, Neuen BL, Arnott C, Simpson R, Oh R, Mahaffey KW, Heerspink HJL. Mediators of the effects of canagliflozin on kidney protection in patients with type 2 diabetes. Kidney Int. 2020; 98:769–777. doi: 10.1016/j.kint.2020.04.051CrossrefMedlineGoogle Scholar
  • 14 Fitchett D, Inzucchi SE, Zinman B, Wanner C, Schumacher M, Schmoor C, Ohneberg K, Ofstad AP, Salsali A, George JT, et al. Mediators of the improvement in heart failure outcomes with empagliflozin in the EMPA‐REG OUTCOME trial. ESC Heart Fail. 2021; 8:4517–4527. doi: 10.1002/ehf2.13615CrossrefMedlineGoogle Scholar
  • 15 Waijer SW, Xie D, Inzucchi SE, Zinman B, Koitka‐Weber A, Mattheus M, von Eynatten M, Inker LA, Wanner C, Heerspink HJL. Short‐term changes in albuminuria and risk of cardiovascular and renal outcomes in type 2 diabetes mellitus: a post hoc analysis of the EMPA‐REG OUTCOME trial. J Am Heart Assoc. 2020; 9:e016976. doi: 10.1161/JAHA.120.016976LinkGoogle Scholar
  • 16 Inzucchi SE, Zinman B, Fitchett D, Wanner C, Ferrannini E, Schumacher M, Schmoor C, Ohneberg K, Johansen OE, George JT, et al. How does empagliflozin reduce cardiovascular mortality? Insights from a mediation analysis of the EMPA‐REG OUTCOME trial. Diabetes Care. 2018; 41:356–363. doi: 10.2337/dc17-1096CrossrefMedlineGoogle Scholar
  • 17 Segar MW, Kolkailah AA, Frederich R, Pong A, Cannon CP, Cosentino F, Dagogo‐Jack S, McGuire DK, Pratley RE, Liu CC, et al. Mediators of ertugliflozin effects on heart failure and kidney outcomes among patients with type 2 diabetes mellitus. Diabetes Obes Metab. 2022; 24:1829–1839. doi: 10.1111/dom.14769CrossrefMedlineGoogle Scholar
  • 18 Lange T, Hansen KW, Sorensen R, Galatius S. Applied mediation analyses: a review and tutorial. Epidemiol Health. 2017; 39:e2017035. doi: 10.4178/epih.e2017035CrossrefMedlineGoogle Scholar
  • 19 Rodriguez‐Valadez JM, Tahsin M, Fleischmann KE, Masharani U, Yeboah J, Park M, Li L, Weber E, Li Y, Berkalieva A, et al. Cardiovascular and renal benefits of novel diabetes drugs by baseline cardiovascular risk: a systematic review, meta‐analysis, and meta‐regression. Diabetes Care. 2023; 46:1300–1310. doi: 10.2337/dc22-0772CrossrefMedlineGoogle Scholar
  • 20 WebPlotDigitizer 4.6 . WebPlotDigitizer. Accessed June 15, 2023. https://automeris.io/WebPlotDigitizerGoogle Scholar
  • 21 VanderWeele TJ, Tchetgen Tchetgen EJ. Mediation analysis with time varying exposures and mediators. J R Stat Soc Series B Stat Methodol. 2017; 79:917–938. doi: 10.1111/rssb.12194CrossrefMedlineGoogle Scholar
  • 22 Vo TT, Vansteelandt S. Challenges in systematic reviews and meta‐analyses of mediation analyses. Am J Epidemiol. 2022; 191:1098–1106. doi: 10.1093/aje/kwac028CrossrefMedlineGoogle Scholar
  • 23 Higgins JP, Altman DG, Gotzsche PC, Juni P, Moher D, Oxman AD, Savovic J, Schulz KF, Weeks L, Sterne JA, et al. The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ. 2011; 343:d5928. doi: 10.1136/bmj.d5928CrossrefMedlineGoogle Scholar
  • 24 Muskiet MHA, Tonneijck L, Huang Y, Liu M, Saremi A, Heerspink HJL, van Raalte DH. Lixisenatide and renal outcomes in patients with type 2 diabetes and acute coronary syndrome: an exploratory analysis of the ELIXA randomised, placebo‐controlled trial. Lancet Diabetes Endocrinol. 2018; 6:859–869. doi: 10.1016/S2213-8587(18)30268-7CrossrefMedlineGoogle Scholar
  • 25 Pfeffer MA, Claggett B, Diaz R, Dickstein K, Gerstein HC, Kober LV, Lawson FC, Ping L, Wei X, Lewis EF, et al. Lixisenatide in patients with type 2 diabetes and acute coronary syndrome. N Engl J Med. 2015; 373:2247–2257. doi: 10.1056/NEJMoa1509225CrossrefMedlineGoogle Scholar
  • 26 Marso SP, Daniels GH, Brown‐Frandsen K, Kristensen P, Mann JF, Nauck MA, Nissen SE, Pocock S, Poulter NR, Ravn LS, et al. Liraglutide and cardiovascular outcomes in type 2 diabetes. N Engl J Med. 2016; 375:311–322. doi: 10.1056/NEJMoa1603827CrossrefMedlineGoogle Scholar
  • 27 Mann JFE, Orsted DD, Brown‐Frandsen K, Marso SP, Poulter NR, Rasmussen S, Tornoe K, Zinman B, Buse JB; LEADER Steering Committee and Investigators. Liraglutide and renal outcomes in type 2 diabetes. N Engl J Med. 2017; 377:839–848. doi: 10.1056/NEJMoa1616011CrossrefMedlineGoogle Scholar
  • 28 Shaman AM, Bain SC, Bakris GL, Buse JB, Idorn T, Mahaffey KW, Mann JFE, Nauck MA, Rasmussen S, Rossing P, et al. Effect of the glucagon‐like peptide‐1 receptor agonists semaglutide and liraglutide on kidney outcomes in patients with type 2 diabetes: pooled analysis of SUSTAIN 6 and LEADER. Circulation. 2022; 145:575–585. doi: 10.1161/CIRCULATIONAHA.121.055459LinkGoogle Scholar
  • 29 Marso SP, Bain SC, Consoli A, Eliaschewitz FG, Jodar E, Leiter LA, Lingvay I, Rosenstock J, Seufert J, Warren ML, et al. Semaglutide and cardiovascular outcomes in patients with type 2 diabetes. N Engl J Med. 2016; 375:1834–1844. doi: 10.1056/NEJMoa1607141CrossrefMedlineGoogle Scholar
  • 30 Mann JFE, Hansen T, Idorn T, Leiter LA, Marso SP, Rossing P, Seufert J, Tadayon S, Vilsboll T. Effects of once‐weekly subcutaneous semaglutide on kidney function and safety in patients with type 2 diabetes: a post‐hoc analysis of the SUSTAIN 1‐7 randomised controlled trials. Lancet Diabetes Endocrinol. 2020; 8:880–893. doi: 10.1016/S2213-8587(20)30313-2CrossrefMedlineGoogle Scholar
  • 31 Holman RR, Bethel MA, Mentz RJ, Thompson VP, Lokhnygina Y, Buse JB, Chan JC, Choi J, Gustavson SM, Iqbal N, et al. Effects of once‐weekly exenatide on cardiovascular outcomes in type 2 diabetes. N Engl J Med. 2017; 377:1228–1239. doi: 10.1056/NEJMoa1612917CrossrefMedlineGoogle Scholar
  • 32 van der Aart‐van der Beek AB, Clegg LE, Penland RC, Boulton DW, Sjostrom CD, Mentz RJ, Holman RR, Heerspink HJL. Effect of once‐weekly exenatide on estimated glomerular filtration rate slope depends on baseline renal risk: a post hoc analysis of the EXSCEL trial. Diabetes Obes Metab. 2020; 22:2493–2498. doi: 10.1111/dom.14175CrossrefMedlineGoogle Scholar
  • 33 Ruff CT, Baron M, Im K, O'Donoghue ML, Fiedorek FT, Sabatine MS. Subcutaneous infusion of exenatide and cardiovascular outcomes in type 2 diabetes: a non‐inferiority randomized controlled trial. Nat Med. 2022; 28:89–95. doi: 10.1038/s41591-021-01584-3CrossrefMedlineGoogle Scholar
  • 34 Hernandez AF, Green JB, Janmohamed S, D'Agostino RB, Granger CB, Jones NP, Leiter LA, Rosenberg AE, Sigmon KN, Somerville MC, et al. Albiglutide and cardiovascular outcomes in patients with type 2 diabetes and cardiovascular disease (Harmony Outcomes): a double‐blind, randomised placebo‐controlled trial. Lancet. 2018; 392:1519–1529. doi: 10.1016/S0140-6736(18)32261-XCrossrefMedlineGoogle Scholar
  • 35 Ferreira JP, Sharma A, Vasques‐Novoa F, Angelico‐Goncalves A, Leite AR, Borges‐Canha M, Carvalho D, Packer M, Zannad F, Leite‐Moreira A, et al. Albiglutide in patients with type 2 diabetes and heart failure: a post‐hoc analysis from Harmony Outcomes. Eur J Heart Fail. 2022; 24:1792–1801. doi: 10.1002/ejhf.2660CrossrefMedlineGoogle Scholar
  • 36 Gerstein HC, Colhoun HM, Dagenais GR, Diaz R, Lakshmanan M, Pais P, Probstfield J, Riesmeyer JS, Riddle MC, Ryden L, et al. Dulaglutide and cardiovascular outcomes in type 2 diabetes (REWIND): a double‐blind, randomised placebo‐controlled trial. Lancet. 2019; 394:121–130. doi: 10.1016/S0140-6736(19)31149-3CrossrefMedlineGoogle Scholar
  • 37 Gerstein HC, Colhoun HM, Dagenais GR, Diaz R, Lakshmanan M, Pais P, Probstfield J, Botros FT, Riddle MC, Ryden L, et al. Dulaglutide and renal outcomes in type 2 diabetes: an exploratory analysis of the REWIND randomised, placebo‐controlled trial. Lancet. 2019; 394:131–138. doi: 10.1016/S0140-6736(19)31150-XCrossrefMedlineGoogle Scholar
  • 38 Husain M, Birkenfeld AL, Donsmark M, Dungan K, Eliaschewitz FG, Franco DR, Jeppesen OK, Lingvay I, Mosenzon O, Pedersen SD, et al. Oral Semaglutide and cardiovascular outcomes in patients with type 2 diabetes. N Engl J Med. 2019; 381:841–851. doi: 10.1056/NEJMoa1901118CrossrefMedlineGoogle Scholar
  • 39 Gerstein HC, Sattar N, Rosenstock J, Ramasundarahettige C, Pratley R, Lopes RD, Lam CSP, Khurmi NS, Heenan L, Del Prato S, et al. Cardiovascular and renal outcomes with efpeglenatide in type 2 diabetes. N Engl J Med. 2021; 385:896–907. doi: 10.1056/NEJMoa2108269CrossrefMedlineGoogle Scholar
  • 40 Lam CSP, Ramasundarahettige C, Branch KRH, Sattar N, Rosenstock J, Pratley R, Del Prato S, Lopes RD, Niemoeller E, Khurmi NS, et al. Efpeglenatide and clinical outcomes with and without concomitant sodium‐glucose cotransporter‐2 inhibition use in type 2 diabetes: exploratory analysis of the AMPLITUDE‐O trial. Circulation. 2022; 145:565–574. doi: 10.1161/CIRCULATIONAHA.121.057934LinkGoogle Scholar
  • 41 Perkovic V, Koitka‐Weber A, Cooper ME, Schernthaner G, Pfarr E, Woerle HJ, von Eynatten M, Wanner C. Choice of endpoint in kidney outcome trials: considerations from the EMPA‐REG OUTCOME® trial. Nephrol Dial Transplant. 2020; 35:2103–2111. doi: 10.1093/ndt/gfz179CrossrefMedlineGoogle Scholar
  • 42 Zinman B, Wanner C, Lachin JM, Fitchett D, Bluhmki E, Hantel S, Mattheus M, Devins T, Johansen OE, Woerle HJ, et al. Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes. N Engl J Med. 2015; 373:2117–2128. doi: 10.1056/NEJMoa1504720CrossrefMedlineGoogle Scholar
  • 43 Yu J, Zhou Z, Mahaffey KW, Matthews DR, Neuen BL, Heerspink HJL, Jardine MJ, Li J, Perkovic V, Neal B, et al. An exploration of the heterogeneity in effects of SGLT2 inhibition on cardiovascular and all‐cause mortality in the EMPA‐REG OUTCOME, CANVAS program, DECLARE‐TIMI 58, and CREDENCE trials. Int J Cardiol. 2021; 324:165–172. doi: 10.1016/j.ijcard.2020.09.050CrossrefMedlineGoogle Scholar
  • 44 Neal B, Perkovic V, Mahaffey KW, de Zeeuw D, Fulcher G, Erondu N, Shaw W, Law G, Desai M, Matthews DR, et al. Canagliflozin and cardiovascular and renal events in type 2 diabetes. N Engl J Med. 2017; 377:644–657. doi: 10.1056/NEJMoa1611925CrossrefMedlineGoogle Scholar
  • 45 Perkovic V, de Zeeuw D, Mahaffey KW, Fulcher G, Erondu N, Shaw W, Barrett TD, Weidner‐Wells M, Deng H, Matthews DR, et al. Canagliflozin and renal outcomes in type 2 diabetes: results from the CANVAS program randomised clinical trials. Lancet Diabetes Endocrinol. 2018; 6:691–704. doi: 10.1016/S2213-8587(18)30141-4CrossrefMedlineGoogle Scholar
  • 46 Li J, Woodward M, Perkovic V, Figtree GA, Heerspink HJL, Mahaffey KW, de Zeeuw D, Vercruysse F, Shaw W, Matthews DR, et al. Mediators of the effects of canagliflozin on heart failure in patients with type 2 diabetes. JACC Heart Fail. 2020; 8:57–66. doi: 10.1016/j.jchf.2019.08.004CrossrefMedlineGoogle Scholar
  • 47 Wiviott SD, Raz I, Bonaca MP, Mosenzon O, Kato ET, Cahn A, Silverman MG, Zelniker TA, Kuder JF, Murphy SA, et al. Dapagliflozin and cardiovascular outcomes in type 2 diabetes. N Engl J Med. 2019; 380:347–357. doi: 10.1056/NEJMoa1812389CrossrefMedlineGoogle Scholar
  • 48 Cahn A, Mosenzon O, Wiviott SD, Rozenberg A, Yanuv I, Goodrich EL, Murphy SA, Bhatt DL, Leiter LA, McGuire DK, et al. Efficacy and safety of dapagliflozin in the elderly: analysis from the DECLARE‐TIMI 58 study. Diabetes Care. 2020; 43:468–475. doi: 10.2337/dc19-1476CrossrefMedlineGoogle Scholar
  • 49 Kolkailah AA, Wiviott SD, Raz I, Murphy SA, Mosenzon O, Bhatt DL, Leiter LA, Wilding JPH, Gause‐Nilsson I, Sabatine MS, et al. Effect of dapagliflozin on hematocrit in patients with type 2 diabetes at high cardiovascular risk: observations from DECLARE‐TIMI 58. Diabetes Care. 2022; 45:e27–e29. doi: 10.2337/dc21-1668CrossrefMedlineGoogle Scholar
  • 50 Perkovic V, Jardine MJ, Neal B, Bompoint S, Heerspink HJL, Charytan DM, Edwards R, Agarwal R, Bakris G, Bull S, et al. Canagliflozin and renal outcomes in type 2 diabetes and nephropathy. N Engl J Med. 2019; 380:2295–2306. doi: 10.1056/NEJMoa1811744CrossrefMedlineGoogle Scholar
  • 51 Jardine M, Zhou Z, Lambers Heerspink HJ, Hockham C, Li Q, Agarwal R, Bakris GL, Cannon CP, Charytan DM, Greene T, et al. Kidney, cardiovascular, and safety outcomes of canagliflozin according to baseline albuminuria: a CREDENCE secondary analysis. Clin J Am Soc Nephrol. 2021; 16:384–395. doi: 10.2215/CJN.15260920CrossrefMedlineGoogle Scholar
  • 52 Oshima M, Neuen BL, Jardine MJ, Bakris G, Edwards R, Levin A, Mahaffey KW, Neal B, Pollock C, Rosenthal N, et al. Effects of canagliflozin on anaemia in patients with type 2 diabetes and chronic kidney disease: a post‐hoc analysis from the CREDENCE trial. Lancet Diabetes Endocrinol. 2020; 8:903–914. doi: 10.1016/S2213-8587(20)30300-4CrossrefMedlineGoogle Scholar
  • 53 Oshima M, Neuen BL, Li J, Perkovic V, Charytan DM, de Zeeuw D, Edwards R, Greene T, Levin A, Mahaffey KW, et al. Early change in albuminuria with canagliflozin predicts kidney and cardiovascular outcomes: a posthoc analysis from the CREDENCE trial. J Am Soc Nephrol. 2020; 31:2925–2936. doi: 10.1681/ASN.2020050723CrossrefMedlineGoogle Scholar
  • 54 Cannon CP, Pratley R, Dagogo‐Jack S, Mancuso J, Huyck S, Masiukiewicz U, Charbonnel B, Frederich R, Gallo S, Cosentino F, et al. Cardiovascular outcomes with ertugliflozin in type 2 diabetes. N Engl J Med. 2020; 383:1425–1435. doi: 10.1056/NEJMoa2004967CrossrefMedlineGoogle Scholar
  • 55 Cherney DZI, Charbonnel B, Cosentino F, Dagogo‐Jack S, McGuire DK, Pratley R, Shih WJ, Frederich R, Maldonado M, Pong A, et al. Effects of ertugliflozin on kidney composite outcomes, renal function and albuminuria in patients with type 2 diabetes mellitus: an analysis from the randomised VERTIS CV trial. Diabetologia. 2021; 64:1256–1267. doi: 10.1007/s00125-021-05407-5CrossrefMedlineGoogle Scholar
  • 56 McMurray JJV, Solomon SD, Inzucchi SE, Kober L, Kosiborod MN, Martinez FA, Ponikowski P, Sabatine MS, Anand IS, Belohlavek J, et al. Dapagliflozin in patients with heart failure and reduced ejection fraction. N Engl J Med. 2019; 381:1995–2008. doi: 10.1056/NEJMoa1911303CrossrefMedlineGoogle Scholar
  • 57 Petrie MC, Verma S, Docherty KF, Inzucchi SE, Anand I, Belohlavek J, Bohm M, Chiang CE, Chopra VK, de Boer RA, et al. Effect of dapagliflozin on worsening heart failure and cardiovascular death in patients with heart failure with and without diabetes. JAMA. 2020; 323:1353–1368. doi: 10.1001/jama.2020.1906CrossrefMedlineGoogle Scholar
  • 58 Docherty KF, Curtain JP, Anand IS, Bengtsson O, Inzucchi SE, Kober L, Kosiborod MN, Langkilde AM, Martinez FA, Ponikowski P, et al. Effect of dapagliflozin on anaemia in DAPA‐HF. Eur J Heart Fail. 2021; 23:617–628. doi: 10.1002/ejhf.2132CrossrefMedlineGoogle Scholar
  • 59 Anker SD, Butler J, Filippatos G, Ferreira JP, Bocchi E, Bohm M, Brunner‐La Rocca HP, Choi DJ, Chopra V, Chuquiure‐Valenzuela E, et al. Empagliflozin in heart failure with a preserved ejection fraction. N Engl J Med. 2021; 385:1451–1461. doi: 10.1056/NEJMoa2107038CrossrefMedlineGoogle Scholar
  • 60 Ferreira JP, Zannad F, Butler J, Filippatos G, Pocock SJ, Brueckmann M, Steubl D, Schueler E, Anker SD, Packer M. Association of empagliflozin treatment with albuminuria levels in patients with heart failure: a secondary analysis of EMPEROR‐pooled. JAMA Cardiol. 2022; 7:1148–1159. doi: 10.1001/jamacardio.2022.2924CrossrefMedlineGoogle Scholar
  • 61 Heerspink HJL, Stefansson BV, Correa‐Rotter R, Chertow GM, Greene T, Hou FF, Mann JFE, McMurray JJV, Lindberg M, Rossing P, et al. Dapagliflozin in patients with chronic kidney disease. N Engl J Med. 2020; 383:1436–1446. doi: 10.1056/NEJMoa2024816CrossrefMedlineGoogle Scholar
  • 62 Persson F, Rossing P, Vart P, Chertow GM, Hou FF, Jongs N, McMurray JJV, Correa‐Rotter R, Bajaj HS, Stefansson BV, et al. Efficacy and safety of dapagliflozin by baseline glycemic status: a prespecified analysis from the DAPA‐CKD trial. Diabetes Care. 2021; 44:1894–1897. doi: 10.2337/dc21-0300CrossrefMedlineGoogle Scholar
  • 63 Jongs N, Greene T, Chertow GM, McMurray JJV, Langkilde AM, Correa‐Rotter R, Rossing P, Sjostrom CD, Stefansson BV, Toto RD, et al. Effect of dapagliflozin on urinary albumin excretion in patients with chronic kidney disease with and without type 2 diabetes: a prespecified analysis from the DAPA‐CKD trial. Lancet Diabetes Endocrinol. 2021; 9:755–766. doi: 10.1016/S2213-8587(21)00243-6CrossrefMedlineGoogle Scholar
  • 64 Packer M, Anker SD, Butler J, Filippatos G, Pocock SJ, Carson P, Januzzi J, Verma S, Tsutsui H, Brueckmann M, et al. Cardiovascular and renal outcomes with empagliflozin in heart failure. N Engl J Med. 2020; 383:1413–1424. doi: 10.1056/NEJMoa2022190CrossrefMedlineGoogle Scholar
  • 65 Anker SD, Butler J, Filippatos G, Khan MS, Marx N, Lam CSP, Schnaidt S, Ofstad AP, Brueckmann M, Jamal W, et al. Effect of empagliflozin on cardiovascular and renal outcomes in patients with heart failure by baseline diabetes status: results from the EMPEROR‐reduced trial. Circulation. 2021; 143:337–349. doi: 10.1161/CIRCULATIONAHA.120.051824LinkGoogle Scholar
  • 66 Bhatt DL, Szarek M, Steg PG, Cannon CP, Leiter LA, McGuire DK, Lewis JB, Riddle MC, Voors AA, Metra M, et al. Sotagliflozin in patients with diabetes and recent worsening heart failure. N Engl J Med. 2021; 384:117–128. doi: 10.1056/NEJMoa2030183CrossrefMedlineGoogle Scholar
  • 67 Bhatt DL, Szarek M, Pitt B, Cannon CP, Leiter LA, McGuire DK, Lewis JB, Riddle MC, Inzucchi SE, Kosiborod MN, et al. Sotagliflozin in patients with diabetes and chronic kidney disease. N Engl J Med. 2021; 384:129–139. doi: 10.1056/NEJMoa2030186CrossrefMedlineGoogle Scholar
  • 68 Solomon SD, McMurray JJV, Claggett B, de Boer RA, DeMets D, Hernandez AF, Inzucchi SE, Kosiborod MN, Lam CSP, Martinez F, et al. Dapagliflozin in heart failure with mildly reduced or preserved ejection fraction. N Engl J Med. 2022; 387:1089–1098. doi: 10.1056/NEJMoa2206286CrossrefMedlineGoogle Scholar
  • 69 McCausland FR, Claggett BL, Vaduganathan M, Desai AS, Jhund P, de Boer RA, Docherty K, Fang J, Hernandez AF, Inzucchi SE, et al. Dapagliflozin and kidney outcomes in patients with heart failure with mildly reduced or preserved ejection fraction: a prespecified analysis of the DELIVER randomized clinical trial. JAMA Cardiol. 2022; 8:56–65. doi: 10.1001/jamacardio.2022.4210CrossrefGoogle Scholar
  • 70 Selvaraj S, Vaduganathan M, Claggett BL, Miao ZM, Fang JC, Vardeny O, Desai AS, Shah SJ, Lam CSP, Martinez FA, et al. Blood pressure and dapagliflozin in heart failure with mildly reduced or preserved ejection fraction: DELIVER. JACC Heart Fail. 2023; 11:76–89. doi: 10.1016/j.jchf.2022.09.002CrossrefMedlineGoogle Scholar
  • 71 The EMPA‐KIDNEY Collaborative Group ; Herrington WG, Staplin N, Wanner C, Green JB, Hauske SJ, Emberson JR, Preiss D, Judge P, Mayne KJ, et al. Empagliflozin in patients with chronic kidney disease. N Engl J Med. 2023; 388:117–127. doi: 10.1056/NEJMoa2204233CrossrefMedlineGoogle Scholar
  • 72 Giugliano D, Maiorino MI, Bellastella G, Chiodini P, Esposito K. Glycemic control, preexisting cardiovascular disease, and risk of major cardiovascular events in patients with type 2 diabetes mellitus: systematic review with meta‐analysis of cardiovascular outcome trials and intensive glucose control trials. J Am Heart Assoc. 2019; 8:e012356. doi: 10.1161/JAHA.119.012356LinkGoogle Scholar
  • 73 ACCORD Study Group ; Buse JB, Bigger JT, Byington RP, Cooper LS, Cushman WC, Friedewald WT, Genuth S, Gerstein HC, Ginsberg HN, et al. Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial: design and methods. Am J Cardiol. 2007; 99:21i–33i. doi: 10.1016/j.amjcard.2007.03.003CrossrefMedlineGoogle Scholar
  • 74 ADVANCE Collaborative Group ; Patel A, MacMahon S, Chalmers J, Neal B, Billot L, Woodward M, Marre M, Cooper M, Glasziou P, et al. Intensive blood glucose control and vascular outcomes in patients with type 2 diabetes. N Engl J Med. 2008; 358:2560–2572. doi: 10.1056/NEJMoa0802987CrossrefMedlineGoogle Scholar
  • 75 Duckworth W, Abraira C, Moritz T, Reda D, Emanuele N, Reaven PD, Zieve FJ, Marks J, Davis SN, Hayward R, et al. Glucose control and vascular complications in veterans with type 2 diabetes. N Engl J Med. 2009; 360:129–139. doi: 10.1056/NEJMoa0808431CrossrefMedlineGoogle Scholar
  • 76 Skyler JS, Bergenstal R, Bonow RO, Buse J, Deedwania P, Gale EA, Howard BV, Kirkman MS, Kosiborod M, Reaven P, et al. Intensive glycemic control and the prevention of cardiovascular events: implications of the ACCORD, ADVANCE, and VA diabetes trials: a position statement of the American Diabetes Association and a scientific statement of the American College of Cardiology Foundation and the American Heart Association. J Am Coll Cardiol. 2009; 53:298–304. doi: 10.1016/j.jacc.2008.10.008CrossrefMedlineGoogle Scholar
  • 77 Hasebe M, Yoshiji S, Keidai Y, Minamino H, Murakami T, Tanaka D, Fujita Y, Harada N, Hamasaki A, Inagaki N. Efficacy of antihyperglycemic therapies on cardiovascular and heart failure outcomes: an updated meta‐analysis and meta‐regression analysis of 35 randomized cardiovascular outcome trials. Cardiovasc Diabetol. 2023; 22:62. doi: 10.1186/s12933-023-01773-zCrossrefMedlineGoogle Scholar
  • 78 Ghosh‐Swaby OR, Goodman SG, Leiter LA, Cheng A, Connelly KA, Fitchett D, Juni P, Farkouh ME, Udell JA. Glucose‐lowering drugs or strategies, atherosclerotic cardiovascular events, and heart failure in people with or at risk of type 2 diabetes: an updated systematic review and meta‐analysis of randomised cardiovascular outcome trials. Lancet Diabetes Endocrinol. 2020; 8:418–435. doi: 10.1016/S2213-8587(20)30038-3CrossrefMedlineGoogle Scholar
  • 79 Ghosal S, Sinha B, Mukherjee R. Heterogeneity in cardiovascular death or hospitalization for heart failure benefits with flozins is linked to weight. ESC Heart Fail. 2023; 10:1242–1249. doi: 10.1002/ehf2.14296CrossrefMedlineGoogle Scholar
  • 80 Yoshiji S, Minamino H, Tanaka D, Yamane S, Harada N, Inagaki N. Effects of glucagon‐like peptide‐1 receptor agonists on cardiovascular and renal outcomes: a meta‐analysis and meta‐regression analysis. Diabetes Obes Metab. 2022; 24:1029–1037. doi: 10.1111/dom.14666CrossrefMedlineGoogle Scholar
  • 81 Imprialos KP, Boutari C, Stavropoulos K, Doumas M, Karagiannis AI. Stroke paradox with SGLT‐2 inhibitors: a play of chance or a viscosity‐mediated reality?J Neurol Neurosurg Psychiatry. 2017; 88:249–253. doi: 10.1136/jnnp-2016-314704CrossrefMedlineGoogle Scholar
  • 82 Heerspink HJL, Kosiborod M, Inzucchi SE, Cherney DZI. Renoprotective effects of sodium‐glucose cotransporter‐2 inhibitors. Kidney Int. 2018; 94:26–39. doi: 10.1016/j.kint.2017.12.027CrossrefMedlineGoogle Scholar
  • 83 Nguyen A, Khafagy R, Hashemy H, Kuo KHM, Roshandel D, Paterson AD, Dash S. Investigating the association between fasting insulin, erythrocytosis and HbA1c through Mendelian randomization and observational analyses. Front Endocrinol (Lausanne). 2023; 14:1146099. doi: 10.3389/fendo.2023.1146099CrossrefMedlineGoogle Scholar
  • 84 Pera MF, Friedlos F, Mills J, Roberts JJ. Inherent sensitivity of cultured human embryonal carcinoma cells to adducts of cis‐diamminedichloroplatinum(II) on DNA. Cancer Res. 1987; 47:6810–6813.MedlineGoogle Scholar
  • 85 Zhou Z, Jardine MJ, Li Q, Neuen BL, Cannon CP, de Zeeuw D, Edwards R, Levin A, Mahaffey KW, Perkovic V, et al. Effect of SGLT2 inhibitors on stroke and atrial fibrillation in diabetic kidney disease: results from the CREDENCE trial and meta‐analysis. Stroke. 2021; 52:1545–1556. doi: 10.1161/STROKEAHA.120.031623LinkGoogle Scholar
  • 86 Tsai WH, Chuang SM, Liu SC, Lee CC, Chien MN, Leung CH, Liu SJ, Shih HM. Effects of SGLT2 inhibitors on stroke and its subtypes in patients with type 2 diabetes: a systematic review and meta‐analysis. Sci Rep. 2021; 11:15364. doi: 10.1038/s41598-021-94945-4CrossrefMedlineGoogle Scholar
  • 87 Coresh J, Heerspink HJL, Sang Y, Matsushita K, Arnlov J, Astor BC, Black C, Brunskill NJ, Carrero JJ, Feldman HI, et al. Change in albuminuria and subsequent risk of end‐stage kidney disease: an individual participant‐level consortium meta‐analysis of observational studies. Lancet Diabetes Endocrinol. 2019; 7:115–127. doi: 10.1016/S2213-8587(18)30313-9CrossrefMedlineGoogle Scholar
  • 88 Heerspink HJL, Greene T, Tighiouart H, Gansevoort RT, Coresh J, Simon AL, Chan TM, Hou FF, Lewis JB, Locatelli F, et al. Change in albuminuria as a surrogate endpoint for progression of kidney disease: a meta‐analysis of treatment effects in randomised clinical trials. Lancet Diabetes Endocrinol. 2019; 7:128–139. doi: 10.1016/S2213-8587(18)30314-0CrossrefMedlineGoogle Scholar
  • 89 Delahoy PJ, Magliano DJ, Webb K, Grobler M, Liew D. The relationship between reduction in low‐density lipoprotein cholesterol by statins and reduction in risk of cardiovascular outcomes: an updated meta‐analysis. Clin Ther. 2009; 31:236–244. doi: 10.1016/j.clinthera.2009.02.017CrossrefMedlineGoogle Scholar
  • 90 Lau J, Terrin N, Fu R. Expanded guidance on selected quantitative synthesis topics. Methods Guide for Effectiveness and Comparative Effectiveness Reviews. AHRQ Publication No. 13‐EHC024‐EF. Agency for Healthcare Research and Quality; March 2013. Accessed June 30, 2023. https://effectivehealthcare.ahrq.gov/products/quantitative‐synthesis‐expanded‐methods/methodsGoogle Scholar

eLetters(0)

eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.