Benefits and harms of sodium‐glucose co‐transporter‐2 inhibitors (SGLT2‐I) and renin–angiotensin–aldosterone system inhibitors (RAAS‐I) versus SGLT2‐Is alone in patients with type 2 diabetes: A systematic review and meta‐analysis of randomized controlled trials

We registered this systematic review and meta‐analysis in the PROSPERO prospective register of systematic reviews (CRD42021251601). It was conducted using a predefined protocol and in accordance with PRISMA guidelines (Appendix). MEDLINE, Embase and the Cochrane Library electronic databases were searched from 2012 (being the year of approval of the first SGLT2 inhibitor (dapagliflozin) in the European Union) to 08 May 2021 with no restriction on language. The computer‐based searches combined terms related to the intervention (eg SGLT2 inhibitor, dapagliflozin, canagliflozin, empagliflozin and ertugliflozin), comparator (eg RAAS inhibitor, ACE‐I), ARB, DRI) and population (eg type 2 diabetes) in humans. A RCT design search filter was employed. Details on the search strategy are provided in Appendix. Titles and abstracts of all initially identified citations were initially screened by one author (SS) to assess their suitability for potential inclusion, followed by the acquisition of full texts for detailed evaluation. Full‐text evaluation was independently conducted by two authors (SS and SKK). The reference lists of key studies and review articles were manually scanned for additional studies. 1 2

Randomized controlled, open or blinded trials that assessed the effects of the combination of SGLT2 and RAAS inhibitors compared with SGLT2 inhibitors in adults with type 2 diabetes and reported on renal or cardiovascular outcomes or adverse events were eligible. Randomized controlled trials that had also compared SGLT2‐I treatment with a placebo or standard care and reported outcomes according to whether patients were receiving RAAS‐Is or not at baseline were considered. We excluded the following: (i) studies that specifically enrolled only patients with known renal insufficiency or established renal parenchymal disease without diabetes mellitus and (ii) studies that recruited patients with a history of diabetic ketoacidosis, type 1 diabetes mellitus, history of hereditary glucose‐galactose malabsorption, primary renal glucosuria or renal disease that required treatment with immunosuppressive agents.

One author (SKK) initially extracted data from eligible studies using a predesigned data collection form and a second author (SS) independently checked the data with that in original articles. A consensus was reached in case of any inconsistency with involvement of a third (KK). Data were extracted on the following: first author, publication year, study year, specific study design, baseline population including duration of years with type 2 diabetes, proportion of men, geographical location, average age at baseline, numbers enrolled and randomized, allocation concealment, blinding, type of SGLT2‐I and dosage; duration of treatment or follow‐up; treatment comparisons; and nature of outcome events and their numbers. We extracted risk estimates when reported.

The primary outcomes were defined as (i) the 3‐point major adverse cardiovascular events (MACE), composite of death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke (composite cardiovascular outcome) and (ii) serum creatinine doubling, initiation of renal replacement therapy (RRT) or death from renal disease (composite renal outcome). Secondary outcomes were (i) cardiovascular death, (ii) heart failure (HF) hospitalization, (iii) composite outcome of cardiovascular death or HF hospitalization, (iv) decline in estimated GFR, (v) RRT, (vi) doubling of serum creatinine level, (vii) other renal and cardiovascular outcomes, (viii) glycaemic measures and haemodynamic and metabolic parameters, and (ix) adverse events.

The risk of bias of each of the included trials was assessed using the Cochrane Collaboration's risk of bias tool.23 This tool evaluates seven possible sources of bias which are random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting and other bias. For each individual component, studies were classified into low, unclear and high risk of bias.

We assessed the quality of the body of evidence on each outcome using the Grading of Recommendations Assessment, Development and Evaluation (GRADEpro) tool (https://gdt.gradepro.org↗), based on study limitations, inconsistency of effect, imprecision, indirectness and publication bias.24 We rated the quality as four levels: high, moderate, low and very low.

Summary measures of association were reported as risk ratios (RRs) with 95% confidence intervals (CIs). Risk ratios were pooled using a fixed effects model given the few studies available for pooling and the absence of substantial heterogeneity across studies. Standard chi‐square tests and the I² statistic were used to quantify the extent of statistical heterogeneity across studies.25, 26 We employed random effects meta‐regression to assess for interactions between RAAS inhibition status and the effect of SGLT2‐Is.27 Only two meta‐analysis could be carried out due to limited data. Given the variety of measures reported for some outcomes and inconsistent reporting by some of the trials, a formal meta‐analysis could not be performed for some of the outcomes. A narrative synthesis was performed for studies that could not be pooled. The findings of such studies were summarized in tables that included the main characteristics of the study and the results in natural units as reported by the investigators. All tests were two‐tailed, and p‐values of 0.05 or less were considered significant. All analyses were conducted using Stata version MP 16 (Stata Corp).

Figure 1 shows the study selection process. The search of relevant databases and manual scanning of reference lists of relevant studies identified 161 potentially relevant citations. After the initial screening of titles and abstracts, 19 articles remained for full text evaluation. Following detailed evaluation, 10 articles were excluded because (i) population was not relevant (n = 7); (ii) duplicate studies (n = 2); and (iii) treatment comparison not relevant (n = 1). The remaining nine articles met the inclusion criteria and were included in the review.28, 29, 30, 31, 32, 33, 34, 35, 36

The nine articles comprised eight studies, of which one was based on a pooled individual patient data (IPD) analysis of 13 trials (Table 1). No RCT specifically investigated the combination of SGLT2 and RAAS inhibitors compared with SGLT2‐Is alone. All eligible studies were based on trials that had investigated the effects of SGLT2 inhibition compared with placebo in patients with type 2 diabetes and reported subgroup results for those treated with or without RAAS‐Is at baseline. Of the 34,551 total participants, 23,109 involved the comparison of SGLT2‐I vs placebo with RAAS inhibition at baseline and 11,442 involved the comparison of SGLT2‐I vs placebo without RAAS inhibition at baseline. In addition to diabetes, patients had other comorbidities such as chronic kidney disease, atherosclerotic CVD or heart failure. Patients had been diagnosed with T2DM and were being managed on standard treatment therapies including ACE‐Is/ARBs, diuretics or calcium channel blockers before inclusion into the trials. All included studies were double‐blinded RCTs. All the studies were conducted in multiple countries. The type of SGLT2‐Is used included dapagliflozin, canagliflozin, empagliflozin, ertugliflozin and sotagliflozin. The majority of trials recruited patients who were at least 18 years old. The average age of participants ranged from 60 to 69 years. The treatment duration ranged from 12 weeks to 6.6 years. Using the Cochrane Collaboration tool, all trials demonstrated low risk of bias in the areas of random sequence generation, allocation concealment, blinding of participants and personnel and incomplete outcome data. Only one trial demonstrated unclear risk of bias for incomplete outcome data and the majority an unclear risk of bias in the areas of selective reporting and other bias (Appendix 3).

Comparing SGLT2‐Is with placebo in those on RAAS‐I treatment at baseline, the RR (95% CIs) for the composite cardiovascular outcome in pooled analysis of three trials was 0.93 (0.85–1.01; I² = 13%; 95% CI 0, 91%; p for heterogeneity = .32). The corresponding risk in those not on RAAS‐I treatment at baseline was 0.78 (0.65–0.93; I² = 0%; 95% CI 0, 90%; p for heterogeneity = .99) (Figure 2). There was no evidence of significant interaction between the effects of SGLT2 inhibition and RAAS inhibition status on the composite cardiovascular outcome (p‐value for meta‐regression = .08).

The composite kidney outcome was reported by only one study. SGLT2‐Is compared with placebo reduced the risk of the composite kidney outcome in those on RAAS inhibition at baseline: 0.52 (95% CI, 0.37–0.74). The corresponding risk for those not treated with RAAS inhibition was 0.65 (95% CI, 0.30–1.39).

In pooled analysis of four trials, the RR (95% CIs) for the composite outcome of cardiovascular death or HF hospitalization was 0.88 (0.76–1.02; I² = 51%; 95% CI 0, 84%; p for heterogeneity = .11) when comparing SGLT2‐Is with placebo in those on RAAS‐I treatment at baseline. For those that were not on RAAS inhibition, the corresponding risk was 0.73 (0.65–0.82; I² = 0%; 95% CI 0, 85%; p for heterogeneity = .50) (Figure 3). There was no evidence of significant interaction between the effects of SGLT2 inhibition and RAAS inhibition status on the composite outcome of cardiovascular death or HF hospitalization (p‐value for meta‐regression = .12).

The outcome of cardiovascular death was reported by one study. SGLT2‐Is compared with placebo reduced the risk of the cardiovascular death in those on RAAS inhibition at baseline: 0.61 (95% CI, 0.48–0.79). The corresponding risk for those not treated with RAAS inhibition was 0.65 (95% CI, 0.39–1.06).

Two studies reported the effect of SGLT2 inhibition vs placebo on estimated GFR changes across the subgroup of RAAS‐I users.29, 35 In the EMPA‐REG OUTCOME trial, the initial change in estimated GFR in patients taking the combination of empagliflozin and RAAS‐Is was higher than for those taking empagliflozin alone (Table 2). The long‐ and post‐treatment changes were also similar for both groups. In the study that pooled data across 13 trials, the effect of dapagliflozin on estimated GFR was similar in patients with or without RAAS inhibition35 (Table 2).

The EMPA‐REG OUTCOME trial reported changes in albuminuria status for the effect of empagliflozin vs placebo across the subgroup of RAAS‐I users.29 Though there appeared to be an improvement in albuminuria status in those with baseline RAAS inhibition than those without (Appendix 4), the report noted that there was no significant evidence of interaction across the subgroup. The effect of dapagliflozin on urinary albumin‐to‐creatinine ratio (UACR) was similar in patients with or without RAAS inhibition in the pooled analysis of 13 trials35 (Table 2).

The EMPA‐REG OUTCOME trial reported outcomes for incident or worsening nephropathy, oedema and acute renal failure for the effect of empagliflozin vs placebo across the subgroup of RAAS‐I users.29 Empagliflozin compared with placebo reduced the risk of the incident or worsening nephropathy, oedema and acute renal failure in those on RAAS inhibition at baseline; the risk was only reduced for oedema in those who were not on RAAS inhibition treatment (Appendix 5).

In the pooled analysis of 13 trials, the effect of dapagliflozin on HbA1c and haematocrit was similar in patients with or without RAAS inhibition; however, mean reductions in body weight, serum uric acid, SBP and diastolic blood pressure (DBP) were more distinct in patients without RAAS inhibition treatment compared with those with RAAS inhibition treatment at baseline35 (Table 2). The effects of SGLT2 inhibition compared with placebo on the risk of hyperkalaemia, hypokalaemia and hypoglycaemia were similar in both groups (Appendix 6).

Comparing SGLT2‐Is with placebo in those on RAAS‐I treatment at baseline, the RR (95% CIs) for volume depletion in pooled analysis of two trials was 1.06 (0.84–1.33). The corresponding risk in those not on RAAS‐I treatment at baseline was 0.82 (0.47–1.46) (Appendix). 7

In pooled analysis of two trials, the effect of SGLT2 inhibition vs placebo on genital infections was similar in patients with or without RAAS inhibition (Appendix). For urinary tract infection (UTI), the effect of SGLT2 inhibition vs placebo on UTI was also similar in patients with or without RAAS inhibition (Appendix). 8 9

Comparing SGLT2‐Is with placebo in those on RAAS‐I treatment at baseline, the RR (95% CIs) for adverse events in pooled analysis of two trials was 0.99 (0.97–1.00). The corresponding risk in those not on RAAS‐I treatment at baseline was 0.98 (0.95–1.02) (Appendix). 10

GRADE ratings for the relevant outcomes are reported in a summary of findings table in Appendix. GRADE quality of the evidence ranged from very low to moderate. 11

In this systematic review and meta‐analysis from available RCTs, we have evaluated the efficacy and safety outcomes in patients with type 2 diabetes comparing the combination of SGLT2 and RAAS inhibitors with SGLT2‐Is alone. This was achieved by investigating the effects of SGLT2 inhibition compared with placebo in people with type 2 diabetes treated with or without RAAS‐Is at baseline. Our findings show SGLT2 inhibition compared with placebo similarly reduced the risk of major cardiovascular outcomes, improved renal parameters (estimated GFR, volume depletion, changes in albuminuria and electrolyte imbalances) and glycaemic measures (HbA1c and hypoglycaemia) and increased the risk of adverse events including genital infections and UTI in both groups of patients with and without RAAS inhibition. The combination of SGLT2 and RAAS inhibition appeared to reduce the risk of the composite renal outcome, cardiovascular death, incident or worsening nephropathy and acute renal failure, but these results were based on single studies. The study that pooled individual patient data from 13 trials showed distinct reductions in body weight, serum uric acid, SBP and DBP for the combination of SGLT2 and RAAS inhibition than SGLT2 inhibition alone. The quality of the evidence ranged from very low to moderate.

A previous pooled analysis of individual level data from 13 placebo‐controlled trials investigating the effects of dapagliflozin on cardio‐renal risk factors in patients with type 2 diabetes with increased albuminuria treated with or without RAAS‐Is at baseline reported similar clinically relevant improvements in metabolic and haemodynamic parameters.35 In a meta‐analysis of 8 RCTs which compared combined therapy of SGLT2‐Is and ACEIs/ARBs with placebo plus ACEIs/ARBs in patients with type 2 diabetes, the combination therapy showed significant reduction in glycaemic parameters, body weight, blood pressure and lower risk of adverse events.21 Another recent meta‐analysis demonstrated that combination therapy with SGLT2‐Is and ACEIs/ARBs compared with ACEIs/ARBs was well‐tolerated and achieved better control of blood pressure, improvement of renal outcomes, alleviation of long‐term renal function and a decrease in blood glucose and body weight, but an increased risk of hypoglycaemia.22 To our knowledge, this is the first aggregate meta‐analysis to attempt to evaluate whether the combination of SGLT2 and RAAS inhibitors provides better cardio‐renal clinical outcomes in patients with type 2 diabetes compared with SGLT2‐Is alone. Our overall results suggest that treatment with SGLT2‐Is provides similar clinical effectiveness and safety in patients with type 2 diabetes treated with or without RAAS inhibition. The combination of SGLT2 and RAAS inhibition may improve some renal outcomes and parameters such as body weight and blood pressure compared to SGLT2 inhibition alone, but further evaluation is needed.

For the past two decades, landmark trials37, 38 have demonstrated that renin‐angiotensin‐aldosterone system blockade is an efficacious method for the protection of both cardiovascular and renal systems. Despite this, there is some residual risk for both cardiovascular and renal outcomes,39 thus necessitating the requirement for further additive treatment options. In our analysis, SGLT2 inhibition compared with placebo reduced the risk of major cardiovascular outcome, but this reduction did not reach statistical significance, as this achievement was expected in a well‐treated population on renin‐angiotensin‐aldosterone system blockade. Thus, in the populations not on RAAS‐Is, the reductions in both composite cardiovascular outcome and composite outcome of cardiovascular death or HF hospitalization were both statistically significant. There is a wealth of clinical data on the renal and cardiovascular protection effects of SGLT2‐Is. They work by targeting target renal tubular glucose reabsorption, thereby exerting glucose lowering effects through glucosuria.40 SGLT2‐Is exert renal protection effects in type 2 diabetes by altering renal haemodynamics, reducing intraglomerular pressure, attenuating diabetes‐associated hyperfiltration and tubular hypertrophy, and reducing the tubular toxicity of glucose. They also reduce albuminuria, serum uric acid without potassium abnormalities, blood pressure, afferent arteriole vasoconstriction, osmotic diuresis, weight loss, and the workload of the proximal tubules to improve tubulointerstitial hypoxia, and then allow fibroblasts to resume normal erythropoietin production.41 The main functions of the RAAS are regulating fluid volume, blood pressure and the vascular response to injury and inflammation.42 Inappropriate activation of the RAAS causes increases in levels of angiotensin II, which lead to end‐organ damage as a result of direct injury to vascular, renal and cardiac tissues. The most commonly used RAAS blockers include ACEIs and ARBs; ACEIs work by reducing the conversion of angiotensin I to angiotensin II, whereas ARBs block the binding of angiotensin II to angiotensin 1 receptor.43 These RAAS blockers are effective for treating systemic hypertension, HF and renal insufficiency.19 Inhibition of the RAAS constitutes the main therapeutic stay in diabetic nephropathy over the last few decades. These RAAS blockers (ACEIs and ARBs) reduce the incidence of progression to end‐stage kidney disease and major adverse cardiovascular outcomes.44

SGLT2 and RAAS inhibitors play different roles at different sites in the kidney, and it has been suggested that their combination might exert synergistic effects on the kidney.41 The vasodilatation effect of RAAS‐Is and natriuretic effect of SGLT2‐Is can also complement each other to reduce systemic oxidative stress and inflammation, which can reduce the incidence of cardiovascular events.45 Several clinical studies have indicated that the combination of SGLT2‐Is with ACEIs/ARBs was associated with greater cardio‐ and reno‐protection and improvement in glycaemic measures, blood pressure and body weight and was well tolerated.21, 22, 46, 47 It may appear that our findings are at odds with the existing evidence, but this is likely because our evaluation which was mainly based on study level subgroup analyses, precluded a head‐to‐head comparison between the combination therapy (SGLT2 plus RAAS inhibitors) and SGLT2‐I. Furthermore, our analysis was limited by the few studies available for pooling. Nevertheless, our findings do suggest that the combination of SGLT2 and RAAS inhibitors may be similar in efficacy and safety if not superior to SGLT2‐Is alone.

Our overall study findings show that the combination of SGLT2 and RAAS inhibitors may have similar cardiovascular and renal benefits in patients with type 2 diabetes compared with SGLT2 inhibitors alone. There is a likelihood that the combination of SGLT2 and RAAS inhibitors may be superior compared to SGLT2‐Is alone in the prevention of renal deterioration in addition to improving body weight and blood pressure, though further data are needed to confirm this. With the rapid increase in the prevalence of type 2 diabetes globally because of increasingly poor lifestyle choices, morbidity and deaths attributable to diabetes will experience a steep increase. A large armamentarium of therapeutic options is urgently needed for the management of type 2 diabetes. For the last two decades, pharmacological inhibition of the RAAS using RAAS‐Is has been the major focus for the management of diabetes nephropathy, which has been associated with good results. The RAAS‐Is have also been used for their cardioprotective effects. Previous studies have shown that combined therapy of SGLT2 and RAAS inhibitors is superior to RAAS‐I therapy alone in patients with type 2 diabetes.21, 22 Taken, the overall results together suggest that SGLT2 inhibition has superior cardio and reno‐protective effects over RAAS inhibition in type 2 diabetes treatment. The use of SGLT2 inhibition as a first line therapy in type 2 diabetes or its early use in the prevention of renal deterioration and cardiovascular complications in addition to its glycaemic control deserves further study. The absence of a significant benefit of the combination of SGLT2‐I and RAAS‐Is on both composite cardiovascular outcome and composite outcome of cardiovascular death or HF hospitalization leaves room for the use of only SGLT2‐Is in populations that may be struggling with polypharmacy and de‐prescribing of some agents necessary. In these situations, the use of only SGLT2‐Is could yield similar outcomes as the combination.

The strengths of the current evaluation deserve consideration. First is the novelty; though a number of RCTs comparing SGLT2‐Is with placebo have reported outcomes among subgroups of patients with or without RAAS inhibition, no review has previously synthesized the evidence. Previous reviews have rather compared combined therapy of SGLT2‐Is and ACEIs/ARBs with ACEIs/ARBs in patients with type 2 diabetes.21, 22 Second, our population of study was clearly defined, which was based on patients with type 2 diabetes treated with or without RAAS inhibition at baseline. Third, our review was prespecified to include only RCTs, which represent the gold standard study designs for evaluating the effectiveness of interventions. Fourth, to minimize selective reporting, we evaluated a comprehensive panel of efficacy and safety outcomes as reported by the individual studies. Finally, we conducted interaction analyses where possible to assess statistical differences in the effect of the two interventions (SGLT2 plus RAAS inhibition vs SGLT2 inhibition). The limitations were inherent and unavoidable. Though we performed quantitative synthesis of the data where possible, inconsistent reporting of outcome measures from some of the studies and findings based on single reports precluded pooling of all available data. Most of the data were based on subgroup analyses reported by the trials, which may be misleading. A head‐to‐head comparison of the two interventions was not possible. Definitive trials that are powered to compare the combination of SGLT2 and RAAS inhibition with SGLT2‐Is alone in people with type 2 diabetes are warranted.

The corresponding author had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. This study is based in data from published articles.