Sodium-glucose co-transporter-2 inhibitors and all-cause mortality: A meta-analysis of randomized controlled trials

Silverii Giovanni Antonio1, Monami Matteo2, Mannucci Edoardo1 1

Abstract:

The present meta-analysis is aimed at assessing the effects of SGLT2 inhibitors on all-cause mortality and differences across different trials and molecules of the class. We included all randomized clinical trials with duration of treatment longer than 52 weeks, enrolling at least 100 patients in each arm, and comparing a SGLT2 inhibitor with any comparator or placebo. Out of 139, 235, and 145 items identified, 21 trials were selected, enrolling 39,593 and 30,771 patients in SGLT2 inhibitor and comparator arms, respectively, with a median duration of 104 weeks, and reporting 2,474 and 2,298 deaths for SGLT2 inhibitors and comparators, respectively. No relevant heterogeneity was found (I2= 17%). Treatment with SGLT2 inhibitors was associated with a significant reduction in all-cause mortality (MH-OR [95 % CI] 0.86 [0.81, 0.91] p<0.00001). Meta-regression analyses found a significant direct association of treatment effect only with proportion of Asian subjects enrolled, and an inverse correlation with the proportion of Caucasian patients. In conclusion, SGLT2 inhibitors reduce all-cause mortality in randomized-controlled trials.

Keywords: SGLT2 inhibitors, meta-analysis, canagliflozin, dapagliflozin, empagliflozin, ertugliflozin

Introduction:

In the EMPAREG-OUTCOME study, empagliflozin was associated with an unexpected 32% reduction in all-cause mortality, as compared to placebo [1]. In other cardiovascular safety trials in type 2 diabetes, canagliflozin [2][3], dapagliflozin [4], and ertugliflozin[5] did not produce a significant reduction in mortality. These results suggested the hypothesis of differences across sodium-glucose transporter 2 (SGLT2) inhibitors on the effects on mortality, although underlying mechanisms remained unknown [6]. Recent evidence from trials performed with dapagliflozin and empagliflozin on patients with heart failure[7] [8] and chronic kidney disease [9] adds relevant information on the effects of SGLT2 inhibitors on mortality. Aim of the present meta-analysis is the assessment of the effects of SGLT2 inhibitors on all-cause mortality, verifying heterogeneity and differences across different trials and molecules of the class.

Methods:

The review protocol was registered on the University of York Centre for Reviews and Dissemination Prospero Web site (registration number: CRD42020193059). An Embase and MEDLINE search for SGLT2 inhibitors available in Europe and USA (canagliflozin, dapagliflozin, empagliflozin, ertugliflozin,) was performed, collecting all randomized clinical trials on humans up to September 16, 2020. The identification of relevant abstracts, the selection of studies based on the criteria described below, and the subsequent data extraction were performed independently by two of the authors (G.A.S., M.M.), and conflicts resolved by the third investigator (E.M.). Completed but still unpublished trials were identified through a search of www.clinicaltrials.gov Web site, using the same keywords, for the identification of further unpublished and otherwise undisclosed trials. A meta-analysis was performed, including all randomized clinical trials with duration of treatment longer than 52 weeks, enrolling at least 100 patients in each arm, and comparing a SGLT2 inhibitor with placebo or any other non-SGLT2 inhibitor drug. Trials with less than 100 patients were excluded, because they would have provided little contribution to the number of events while increasing heterogeneity, thus reducing the reliability of results. We included only trials with molecules and doses approved by FDA, EMA, or both (dapagliflozin 10mg/day, empagliflozin 10 and 25mg/day, canagliflozin 100 and 300mg/day). Results of trials were retrieved from the primary publication and, if needed, from other publications referring to the same trial. When information was unavailable on published papers, data were retrieved from www.clinicaltrials.gov study. Retrieved data included mortality, the main features of each trial, such as concurrent therapy, principal endpoint, estimated glomerular filtration rate (eGFR), and baseline characteristics of enrolled patients: (age, BMI, SBP, eGFR, HbA1c). The quality of trials was assessed using the Cochrane Collaboration’s tool for assessing risk of bias in randomized controlled trials; quality was not used as a criterion for the selection of trials, but only for descriptive purposes. The outcome of this analysis was the effect of SGLT2 inhibitors, compared with placebo or other active drugs, on all-cause mortality. Heterogeneity was assessed by using I2 statistics. In order to estimate possible publication or disclosure bias, we used funnel plots and the Begg’s adjusted rank correlation test, including published and unpublished, but disclosed, trials. A fixed-effects model was applied, calculating Mantel–Haenszel odds ratio with 95 % confidence interval (MH-OR) for all-cause mortality, on an intention-to-treat basis. Sensitivity analyses were performed using random-effects model, and excluding trials with less than 100 events. Subgroup analysis was performed to explore differences in mortality between molecules, whereas a metaregression log odds ratio (95% CI) analysis was performed to assess the correlation of trials characteristics with the difference in all-cause mortality between SGLT2 inhibitors and placebo. All analyses specified above were performed using Comprehensive Meta-Analysis version 3, Biostat (Englewood, NJ, USA), and Review Manager (RevMan) version 5.4 (The Cochrane Collaboration, 2020). This meta-analysis was reported following the PRISMA checklist [10].

Results:

Out of 139, 235, and 145 items identified through Embase, MEDLINE, and www.clinicaltrials.gov, respectively, 21 trials were selected, as shown in Fig. 1S of supplementary materials. The quality of trials was generally good (Table 1S of supplementary materials). The trials fulfilling the inclusion criteria enrolled 39,593 and 30,771 patients in SGLT2 inhibitor and comparator arms, respectively, with a median duration of treatment of 104 weeks. The main characteristics of the selected trials are reported in Tables 1s of supplementary materials. Of the 21 trials fulfilling the inclusion criteria, one reported no deaths in both treatment arms [11], and was not therefore included in the analysis. The number of reported deaths was 2,474 and 2,298 for SGLT2 inhibitors and comparators, respectively. Funnel plot analysis (Fig. 2S of supplementary materials) and Kendall’s tau (-0.21; p = 0.19) did not suggest any relevant publication bias.
Treatment with SGLT2 inhibitors was associated with a significant reduction in all-cause mortality (MH-OR [95 % CI] 0.86 [0.81, 0.91], p<0.00001; Fig. 1), with low heterogeneity (I2=17%). Similar results were obtained in sensitivity analyses with a random-effects model (MH-OR [95 % CI] 0.85 [0.80, 0.91], p<0.00001; I2=17%), and in a further analysis excluding trials with less than 100 events (MH-OR [95 % CI] 0.85 [0.80, 0.91], p<0.00001; I2=46%) The effect on mortality was statistically significant for canagliflozin, dapagliflozin and empagliflozin, but not for ertugliflozin; however, the difference across molecules was not statistically significant (Fig. 1).
Despite the relatively low heterogeneity, meta-regression analyses were performed in order to explore possible determinants of the effects of SGLT2 inhibitors on mortality in individual trials. A significant direct association of treatment effect was found with the proportion of Asian subjects enrolled (Log OR 95% CI 0.008 [0.0005;0.02], p= 0.037; Table 1; Fig. 3Sa of supplementary materials), whereas an inverse correlation was observed with the proportion of Caucasian patients (Log OR 95% CI -0.01 [-0.002;-0.02], p=0.016; Table 1; Fig. 3SC of supplementary materials). No correlation was observed between treatment effect and mean trial duration, baseline characteristics (mean age, HbA1c, SBP, eGFR, BMI), drug dose, proportion of females enrolled, proportion of African or African-American enrolled, yearly incident mortality in controls, difference in HbA1c from placebo at endpoint, HbA1c reduction from baseline in SGLT2 inhibitors arm, proportion of patients with previous cardiovascular events, proportion of cardiovascular deaths on overall mortality (Table 1, Fig. 3SB of supplementary materials).

Discussion:

In available longer-term randomized trials, SGLT2 inhibitors significantly reduce all-cause mortality. This result is consistent with that of earlier meta-analyses ([12]–[15]), which did not include many of the recent large-scale trials, and which did not allow reliable comparisons across molecules of the class. Since differences across molecules are not statistically significant, the reduction of mortality appears to be a class effect, although differences across molecules cannot be entirely excluded. All-cause mortality was a secondary endpoint in all available trials, which were designed for other, and more frequent, outcomes; as a consequence, the sample size of individual trials is insufficient for a reliable estimation of treatment effect. Considering the relatively low statistical power of each trial for all-cause mortality, differences across studies could be a play of chance. In fact, the calculated heterogeneity across trials was relatively low. A previous analysis limited to the four available cardiovascular outcome trials in type 2 diabetes had revealed a significant heterogeneity, although meta-regression failed to identify its determinants[16]; however, the number of included trials could have been too small to draw reliable conclusions. Despite the fact that heterogeneity was not significant in our analysis, we explored the potential impact on mortality results of several putative moderators, through meta-regression analyses. A greater reduction of mortality was observed in trials enrolling higher proportion of Asian subjects and lower proportion of Caucasians. Notably, a previous meta-analysis reported no difference in efficacy of SGLT2 inhibitors on HbA1c between Asian and non-Asian diabetic subjects [17]; however, the effects of SGLT2 inhibitors on mortality, which have been observed also in trials including non-diabetic subjects[7], [9], could be largely mediated by factors other than glycemic control. In fact, no association was observed in the present analysis between SGLT2 inhibitors effect on mortality and variations of HbA1c from baseline. All other putative determinants of treatment effects, including gender, body mass index, glycemic control and blood pressure, did not appear to be significant moderators of reduction of mortality. Renal function, expressed as mean eGFR at enrolment, also failed to show a significant association with treatment effects on all-cause mortality. Some limitations in the present analysis should be recognized. Meta-regression can generate interesting hypotheses on moderators of treatment effect, but it does not provide definitive evidence. In addition, a relatively small number of large-scale trials accounts for the majority of the results presented in the present paper. Furthermore, some possible moderators of treatment effect could not be explored because not commonly reported. Moreover, it is possible that effects on mortality due to cerebrovascular accidents were different from those on cardiac death, but available information is not detailed enough to explore this point. In conclusion, SGLT2 inhibitors reduce all-cause mortality in randomized-controlled trials, with no apparent differences across molecules of the class. Based on available data, it is not yet possible to clearly identify sub-populations in which the reduction of mortality is more or less pronounced.
 
References
1. B. Zinman et al., “Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes,” N. Engl. J. Med., vol. 373, no. 22, pp. 2117–2128, 2015.
2. B. Neal et al., “Canagliflozin and Cardiovascular and Renal Events in Type 2 Diabetes.,” N. Engl. J. Med., vol. 377, no. 7, pp. 644–657, Aug. 2017.
3. V. Perkovic et al., “Canagliflozin and Renal Outcomes in Type 2 Diabetes and Nephropathy.,” N. Engl.
4. J. Med., vol. 380, no. 24, pp. 2295–2306, Jun. 2019.
5. S. D. Wiviott et al., “Dapagliflozin and cardiovascular outcomes in type 2 diabetes,” N. Engl. J. Med., vol. 380, no. 4, pp. 347–357, 2019.
6. C.P. Cannon et al., “Cardiovascular Outcomes with Ertugliflozin in Type 2 Diabetes.,” N. Engl. J. Med., pp. 1–11, 2020.
7. R. E. Gilbert and K. A. Connelly, “Understanding EMPA-REG OUTCOME.,” The lancet. Diabetes & endocrinology, vol. 3, no. 12. England, pp. 930–931, Dec-2015.
8. J. J. V. McMurray et al., “Dapagliflozin in patients with heart failure and reduced ejection fraction,”
9. N. Engl. J. Med., vol. 381, no. 21, pp. 1995–2008, 2019.
10. M. Packer et al., “Cardiovascular and Renal Outcomes with Empagliflozin in Heart Failure,” N. Engl. J. Med., 2020.
11. H. J. L. Heerspink et al., “Dapagliflozin in Patients with Chronic Kidney Disease.,” N. Engl. J. Med., pp. 1–11, 2020.
12. D.Moher et al., “Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement,” PLoS Med., vol. 6, no. 7, 2009.
13. L.Merker et al., “Empagliflozin as add-on to metformin in people with Type 2 diabetes,” Diabet. Med., vol. 32, no. 12, pp. 1555–1567, 2015.
14. C.-Y. Zou, X.-K. Liu, Y.-Q. Sang, B. Wang, and J. Liang, “Effects of SGLT2 inhibitors on cardiovascular outcomes and mortality in type 2 diabetes: A meta-analysis.,” Medicine (Baltimore)., vol. 98, no. 49, p. e18245, Dec. 2019.
15. M.Monami, I. Dicembrini, and E. Mannucci, “Effects of SGLT-2 inhibitors on mortality and cardiovascular events: a comprehensive meta-analysis of randomized controlled trials.,” Acta Diabetol., vol. 54, no. 1, pp. 19–36, Jan. 2017.
16. H. Tang, Z. Fang, T. Wang, W. Cui, S. Zhai, and Y. Song, “Meta-Analysis of Effects of Sodium-Glucose Cotransporter 2 Inhibitors on Cardiovascular Outcomes and All-Cause Mortality Among Patients With Type 2 Diabetes Mellitus.,” Am. J. Cardiol., vol. 118, no. 11, pp. 1774–1780, Dec. 2016.
17. S. L. Zheng et al., “Association Between Use of Sodium-Glucose Cotransporter 2 Inhibitors,
18. Glucagon-like Peptide 1 Agonists, and Dipeptidyl Peptidase 4 Inhibitors With All-Cause Mortality in Patients With Type 2 Diabetes: A Systematic Review and Meta-analysis.,” JAMA, vol. 319, no. 15, pp. 1580–1591, Apr. 2018.
19. J. Yu 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., Sep. 2020.
20. A. J. Scheen, “SGLT2 Inhibitors as Add-On Therapy to Metformin for People with Type 2 Diabetes: A Review of Placebo-Controlled Trials in Asian versus Non-Asian Patients.,” Diabetes. Metab. Syndr. Obes., vol. 13, pp. 2765–2779, 2020.