The effect of renin–angiotensin–aldosterone system inhibitors on continuous and binary kidney outcomes in subgroups of patients with diabetes: a meta-analysis of randomized clinical trials

Diabetic nephropathy, a complication of diabetes, is the leading cause of kidney failure, responsible for approximately 40% of incident cases [1].

Diabetic nephropathy is characterized by hypertension, variable levels of albuminuria and a progressive loss of kidney function [2, 3]. The progression of histological and pathological changes in diabetic nephropathy are due to hyperglycemia [4]. The histological and pathological changes differ between type 1 and type 2 diabetes (T1DM, T2DM, respectively). In T1DM, hyperglycemia starts earlier hence it causes pure diabetic glomerulopathy that could be evaluated at the stage of microalbuminuria. Whereas in T2DM hyperglycemia starts later in life when kidneys were already damaged due to the long-term effects of many possible promoters of kidney injury such as aging, hypertension, and dyslipidemia. Therefore, there is a heterogenous combination of pathophysiological pathways that sustain structural changes in the kidneys of T2DM patients. Regardless of the involved mechanism, the final common pathway of diabetic nephropathy is kidney fibrosis that is caused by kidney hemodynamic and ischemic abnormalities, oxidative stress and the overactivation of the renin-angiotensin aldosterone system (RAAS) [5, 6]. Clinical practice guidelines recommend prescribing angiotensin converting enzyme (ACE) inhibitors and angiotensin receptor blockers (ARBs) which are the two major classes of renin–angiotensin aldosterone system inhibitors (RAASi) to prevent and manage diabetic nephropathy at any stage [2, 3]. The blockade of RAAS is critical to control blood volume, systemic vascular resistance and electrolyte balance [7, 8]. This results in RAASi protecting the kidneys from developing diabetic nephropathy, as well as slowing the progression of the disease [9, 10]. Therefore, these RAASi are the antihypertensive class of choice recommended for the management of patients with hypertension and DM [2, 3].

A number of meta-analyses have been published on the role of RAASi in renoprotection for patients with diabetes. The authors concluded that ACE inhibitors and ARBs are equally effective in slowing the progression of diabetic nephropathy [9‐12]. However, these meta-analyses of RAASi have focused on cardiovascular and kidney outcomes, and had restricted inclusion and exclusion criteria of eligible studies and limited the included clinical trials to patient populations with one type of diabetes, a specific level of albuminuria, and/or excluding patients with advanced stages of chronic kidney disease (CKD). These findings are therefore applicable to a narrow population, and may be limited in their ability to guide clinical care and decision making for a large proportion of patients with diabetes. In addition, most of the previous meta-analyses evaluated RAASi effect on binary kidney outcomes (e.g. kidney failure, progression to albuminuria, mortality), and seldom provided an analysis of RAASi effect on continuous kidney outcomes (e.g. creatinine clearance level, albuminuria level). To address this knowledge gap we have conducted a systematic review (SR) with broader inclusion criteria to allow for conducting sub-group analysis for different variables and hence to identify their effects on kidney and other health outcomes. The objective of this systematic review and meta-analysis is to compare the effects of ACE inhibitors/ARBs with placebo and other antihypertensives in adults with diabetes on both continuous and binary kidney outcomes.

For this systematic review, we followed the reporting guidance provided in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (Supp 1 Table 1) [13]. The SR protocol was registered and published with PROSPERO (CRD42020149133). A brief summary of the methodology is described here, and is based on PRISMA guidelines for reporting SRs [14].

In this meta-analysis, evidence was drawn from 26,551 patients with diabetes from 46 studies on the effect of RAASi on continuous and binary kidney outcomes. This meta-analysis provided interesting findings on the effects of RAASi on some important continuous outcomes in comparison with placebo or other antihypertensives. RAASi were better than placebo in reducing SrCr and albuminuria. Estimated GFR was slightly increased by RAASi compared to placebo after performing a sensitivity analysis. When compared to active treatments, RAASi resulted in a modest increase of SrCr, led to eGFR decline, and resulted in modest reduction of albuminuria levels. Our analysis shows that the RAASi class was superior to placebo in reducing the risks of kidney failure and doubling of SrCr levels, but not in promoting the regression of albuminuria. RAASi, however, were not superior to other antihypertensive agents in reducing the risks of these kidney outcomes or all-cause mortality. Despite some key differences in the selection criteria (Supp 1 Table 6), the findings of our meta-analysis are consistent with previously published meta-analyses.

What distinguishes our meta-analyses from earlier reviews is the inclusion criteria of clinical trials, as well as our analysis of continuous kidney outcomes.(Supp 1 Table 6) We excluded RCTs with sample size of less than 50 participants to exclude small-size effects on the analysis. We performed subgroup analysis based on sample size to further isolate small-size effect of the studies of less than 100 participants. A duration of 12 months or more was a key inclusion criterion to help study the long-term effects of RAASi. Unlike some earlier meta-analyses, we did not limit our analysis to one type of diabetes or to a specific degree of kidney function or albuminuria, which allowed us to perform a variety of subgroup analyses.

This meta-analysis provided noteworthy findings about the effect of RAASi in each subgroup of patients. We have studied RAASi effects on each kidney outcome in different subgroups of study participants. The analysis showed that specific subgroups of patients had better outcomes with RAASi. Patients with T2DM, macroalbuminuria and an average duration of diabetes more than or equal 10 years had less risk to develop kidney failure in placebo-controlled trials, while longer duration of diabetes, normal kidney function, and hypertension increased the probability to achieve regression of albuminuria in active-controlled trials. Type 1 diabetes and hypertensive patients had higher rates of regression of albuminuria in placebo-controlled trials. As these findings suggest, the type and the duration of diabetes as well as other characteristics can influence the response to interventions on some kidney outcomes, which highlights the importance to consider each patient’s medical history when deciding on starting a treatment for them. These findings point to the need to direct more research initiatives on exploring patients’ characteristics that can predict who would benefit most from each intervention, including the broadening of inclusion criteria in studies, and conducting studies powered to look at divergent subgroups. The interpretation of subgroup analyses should be performed with caution, due to some inherited limitations, majorly because they are observational in nature although being derived from randomized trials [60].

The latest version of the Kidney Disease Improving Global Outcomes (KDIGO) clinical practice guidelines for diabetes management in CKD patients [61] were published in late 2020, and it recommends using RAASi as first-line antihypertensives for patients with diabetes, hypertension and albuminuria. These recommendations were based on four placebo-controlled landmark trials of ARBs on patients with different levels of albuminuria [28, 29, 52, 62]. This recommendation is concordant with those of other guidelines, [28, 29, 62] yet it raises some doubts about the robustness of the evidence behind it. The KDIGO guideline supports its recommendation with evidence from trials of ARBs only. Nevertheless, our analysis on continuous outcomes provides moderate evidence on RAASi’s ability to reduce albuminuria levels more than active-treatments. Cativo et al. [63] reached a similar conclusion, and highlighted that although the effect is statistically significant, the clinical effect is small. In summary, the evidence behind promoting RAASi as the leading class in protecting the diabetic kidneys may not be as robust as commonly believed. The findings of this study suggest that the most important factor for preventing and managing diabetic nephropathy is lowering BP levels, which could be of higher significance than the class of the antihypertensive used to lower BP.

Protecting the diabetic kidneys is not exclusive to antihypertensives, as some novel classes claimed their positions in the competition towards protecting diabetic patients from kidney disease using different mechanisms. For example, the mineralocorticoid receptor antagonist (finerenone) is being evaluated in a large RCT (FIDELIO-DKD), with some preliminary promising results [64]. The new antihyperglycemic agents from the sodium‐glucose co‐transporter‐2 (SGLT2) inhibitors class have also shown protective effects against progression of CKD, with reductions of mortality rates when used in combination with RAASi [65]. Sacubitril/valsartan have shown preservation effects of kidney function in older patients with heart failure, and its role in the management of diabetic nephropathy is to be evaluated [66]. Taking the collective adverse events of these agents into consideration, the prescriber today has more options to consider to reduce the progression of diabetic kidney disease. A prescription that combines these agents with a proper antihypertensive could be viewed as the recipe of kidney protection in patients with diabetes. Nevertheless, more research studies need to be carried out to prove the safety and efficacy of such combinations.

This meta-analysis sheds light on the full spectrum of RAASi effects on kidney outcomes in patients with diabetes, by studying its efficacy on both continuous and binary outcomes. It provides a comprehensive review of the class effect on several subgroups of study participants, which was facilitated by the broad inclusion criteria. While we were attempting to answer a research question on RAASi comparative efficacy, our study raised a challenging question on the role of RAASi in preventing and managing diabetic nephropathy and whether it deserves its place as a first-line therapy in the clinical practice guidelines. The analysis' protocol was initially designed to include studies that reported other relevant kidney outcomes including urinary albumin concentration, albumin creatinine ratio, fractional albumin excretion, and kidney deaths. Therefore, the initial number of included studies were 53 trials. However, there was a very limited number of studies that collectively reported these outcomes, which lead to insufficient data to perform meaningful analysis of these outcomes. Therefore, the final number of included studied was 46.

A few limitations of this meta-analysis should be considered when interpreting and applying its findings. The analysis of the continuous outcomes was performed using a number of statistical assumptions.(Supp 1 Table 4) Another limitation is the degree of heterogeneity between the RCTs that were used to determine the change in the continuous outcomes. These RCTs were published across more than 20 years with variable methodological approaches and reporting qualities that resulted in methodological heterogeneity. The included studies shared a wide range of participant characteristics due to our broad inclusion criteria which resulted in population heterogeneity. The performance of subgroup analysis and sensitivity analysis helped mitigate the effect of this type of heterogeneity. The mean follow-up of included studies ranged between 12 and 72 months. Therefore, we conducted subgroup analyses for each outcome to account the difference in the duration of follow-up between studies. (Supp 2, tables 3–18).

It is noteworthy to mention that most of the RCTs were not powered to detect the changes in the continuous kidney outcomes because these were not primary outcomes. We have calculated the effect size from the reported events numbers because of the heterogeneity in reporting effect sizes between studies. We did not analyze data presented in composite outcomes, because of the inconsistency of the trials in reporting these outcomes as the same composite. However, we analyzed data for each single outcome separately.

This systematic review and meta-analysis identified 46 studies, andshowed that RAASi class was better than placebo in reducing SrCr and albuminuria levels. When compared to other active treatments, RAASi did not reduce SrCr levels, caused a non-significant reduction of eGFR, and resulted in modest reduction of albuminuria levels. These results were reported with considerable statistical heterogeneity. As for binary outcomes, RAASi were superior to placebo but not the other antihypertensive agents in reducing the risks of kidney failure and doubling of SrCr..While our findings revealed the non-superiority of RAASi over other antihypertensives it raised some doubts about the robustness of evidence behind placing RAASi as first-line therapy in managing diabetic nephropathy.