The Discipline of Adding — 16 Trials, and How Heart, Kidney and Metabolism Became One Arena
If a drug is working, why not make it work harder? That idea was decisively refuted in 2013, by VA NEPHRON-D. From then on, the logic of add-on therapy shifted — not "push the same pathway harder" but "add another pathway entirely." From RENAAL, through SGLT2 inhibitors, non-steroidal MRAs and GLP-1 receptor agonists, to STRONG-HF and CONFIDENCE. Lining up 16 trials reveals the path by which a kidney trial's result became a Class I recommendation in heart failure guidelines — and makes visible what has actually gone down, and what has not.
Guidelines this column draws on
The text has been checked against the statements below. Please read the originals, and the current editions, before acting on them.
-
Japanese Circulation Society / Japanese Heart Failure SocietyRegardless of the presence or absence of diabetes, administer an SGLT2 inhibitor (empagliflozin, dapagliflozin) to symptomatic patients with HFrEF for the purpose of reducing cardiovascular death and heart failure hospitalisation (Class I / Level of Evidence A).
-
Japanese Circulation Society / Japanese Heart Failure SocietyIn patients with type 2 diabetes and chronic kidney disease, use an SGLT2 inhibitor to prevent the onset of heart failure or cardiovascular death (Class I / Level A). In patients with type 2 diabetes and chronic kidney disease, use finerenone to prevent the onset of heart failure (Class I / Level A).
-
Japanese Circulation Society / Japanese Heart Failure SocietyFor symptomatic HFpEF/HFmrEF, administer an SGLT2 inhibitor (empagliflozin or dapagliflozin) for the purpose of reducing cardiovascular death or heart failure hospitalisation (Class I / Level A). For symptomatic HFpEF/HFmrEF, consider administering an MRA (finerenone) for the purpose of reducing cardiovascular death or worsening heart failure events (Class IIa / Level B-R).
-
Japanese Society of NephrologyCQ4-5: In patients with diabetic kidney disease (DKD), administration of an SGLT2 inhibitor is recommended, as improvement of renal prognosis and suppression of cardiovascular disease onset are expected [1A]. CQ11-3: In patients with CKD without diabetes, if proteinuria is present, administration of an SGLT2 inhibitor is recommended [1B] (there is no evidence for initiation in the absence of proteinuria or at an eGFR below 20 mL/min/1.73m² [None, D]). CQ4-2: The use of a mineralocorticoid receptor antagonist is suggested, as it may show improvement in urinary albumin in patients with DKD [2C].
-
Japanese Society of NephrologyThe recommendation grades/evidence levels in CKD stages G4 and G5 are as follows: ACE inhibitor/ARB 2C, beta-blocker 2B, MRA None C, SGLT2 inhibitor 2C, ARNI 2C, ivabradine None D.
-
Kidney Disease: Improving Global Outcomes (KDIGO)3.6.4: Regardless of the presence or absence of diabetes, we recommend avoiding any combination of ACE inhibitors, ARBs, and direct renin inhibitors in patients with CKD (1B). 3.7.2: We recommend treatment with an SGLT2 inhibitor in adult patients with CKD who have an eGFR of 20 mL/min/1.73m² or greater and a urine ACR of 200 mg/g or greater, or who have heart failure (regardless of the degree of albuminuria) (1A). 3.8.1: We suggest a non-steroidal MRA with proven renal or cardiovascular benefit in adults with type 2 diabetes, an eGFR greater than 25 mL/min/1.73m², normal serum potassium, and albuminuria greater than 30 mg/g that persists despite maximum tolerated dose of a RAS inhibitor (2A).
-
Kidney Disease: Improving Global Outcomes (KDIGO)Administer RAS inhibitors at the highest tolerated, approved dose. Check blood pressure, serum creatinine, and serum potassium within 2 to 4 weeks of initiation or dose increase. Hyperkalaemia associated with RAS inhibitors can often be managed by measures to lower serum potassium rather than by dose reduction or discontinuation. Continue ACE inhibitors/ARBs unless serum creatinine rises by more than 30% within 4 weeks of initiation or dose increase, and continue even if eGFR falls below 30 mL/min/1.73m². SGLT2 inhibitors may be continued even if eGFR falls below 20 mL/min/1.73m² after initiation, unless not tolerated or renal replacement therapy is started. During prolonged fasting, surgery, or serious acute illness, temporary discontinuation is reasonable in view of the risk of ketosis. A reversible decline in eGFR at initiation is not usually an indication for discontinuation.
The 16 trials behind it
RENAAL Trial: The Classic Foundational Trial Showing ARBs Slow Progression of Type 2 Diabetic Nephropathy N Engl J Med, 2001 IDNT trial: Irbesartan protected against diabetic nephropathy beyond calcium-channel blockade N Engl J Med, 2001 VA NEPHRON-D trial: combining ACE inhibitors and ARBs brings harm without benefit N Engl J Med, 2013 EMPA-REG OUTCOME trial: the first diabetes drug to reduce cardiovascular death N Engl J Med, 2015 CREDENCE trial: the first SGLT2 inhibitor trial with renal outcomes as the primary aim N Engl J Med, 2019 DAPA-CKD Trial: Can SGLT2 Inhibitors Halt CKD Progression Regardless of Diabetes Status? N Engl J Med, 2020 EMPA-KIDNEY trial: does empagliflozin slow renal progression in the broadest CKD population studied? N Engl J Med, 2023 PARADIGM-HF: Does ARNI surpass the ACE inhibitor "standard" (HFrEF)? N Engl J Med, 2014 EMPEROR-Reduced: Does a Second SGLT2 Inhibitor Reproduce the Benefit in HFrEF? N Engl J Med, 2020 DELIVER trial: does dapagliflozin also work in heart failure with preserved/mildly reduced EF? N Engl J Med, 2022 FIDELIO-DKD試験：非ステロイド型MRAフィネレノンは糖尿病性腎症を守れるのか N Engl J Med, 2020 FIGARO-DKD trial: does finerenone reduce cardiovascular events even in "earlier-stage CKD"? N Engl J Med, 2021 FINEARTS-HF: Does the nonsteroidal MRA finerenone work in HFmrEF/HFpEF? N Engl J Med, 2024 FLOW trial: can the GLP-1 receptor agonist semaglutide protect the kidney in "diabetic CKD"? N Engl J Med, 2024 STRONG-HF: Does “fast and certain” optimisation of medication after discharge for acute heart failure change outcomes? Lancet, 2022 Simultaneous combination of finerenone and empagliflozin (CONFIDENCE) N Engl J MedFrom "push the same pathway harder" to "add another pathway entirely"
If a drug is working, why not make it work harder? If RAS inhibitors protect the kidney, then combining an ACE inhibitor with an ARB should protect it more. And indeed, proteinuria did fall further with combination therapy.
But in 2013 this idea was decisively refuted. What refuted it was not lack of efficacy, but harm. From then on, the logic of add-on therapy shifted — not "push the same pathway harder" but "add another pathway entirely."
What has happened over these two-plus decades is not that the number of drugs increased. The question itself was replaced. Instead of "which organ's disease is this," the question became "which mechanism, how many, and when do you add it." As a result, a drug that emerged from a diabetes safety trial ended up as the top-tier recommendation in heart failure guidelines.
This essay lines up 16 trials that have been fact-checked on this site, arranged into four acts, and closes by touching on the newest question of all — how many pathways, in what order, and by when.
Act One: The First Pillar Is Planted in the Kidney
In 2001, two trials appeared side by side in the same issue of NEJM: RENAAL and IDNT. Both tested an ARB in type 2 diabetic nephropathy.
In RENAAL (1,513 patients, mean follow-up 3.4 years), losartan reduced the primary composite (doubling of serum creatinine, end-stage renal disease, or death) by a relative 16% (43.5% vs 47.1%, P=0.02). Broken down: a 25% reduction in doubling of serum creatinine (P=0.006), and a 28% reduction in end-stage renal disease (P=0.002). First hospitalisation for heart failure fell by 32% (P=0.005) — even at this stage, a kidney drug was already reducing heart failure admissions. There was, however, no significant difference in all-cause mortality.
IDNT (1,715 patients, mean 2.6 years) was cleverly designed on the comparator side. With irbesartan, amlodipine and placebo groups matched for target blood pressure, the primary composite was reduced by 20% versus placebo (P=0.02) and by 23% versus amlodipine (P=0.006). Doubling of serum creatinine fell by 33% versus placebo (P=0.003). Despite equivalent blood pressure lowering, the calcium channel blocker conferred no renal protection.
What was established here is that renal protection is not simply a byproduct of blood pressure lowering. And the limitation was fixed at the same moment: neither trial reduced all-cause mortality.
The Turning Point: The First Failure of Addition
Once one pillar stands, the next question naturally becomes "what about a second?" Both ACE inhibitors and ARBs protect the kidney as monotherapy. Combine them, and proteinuria falls even further. Surely, then, the effect would be greater still.
VA NEPHRON-D (2013, 1,448 patients) tested this directly, adding lisinopril to losartan 100mg. The primary composite (decline in eGFR, end-stage renal disease, or death) did not improve, nor did all-cause mortality. Meanwhile, hyperkalaemia occurred at 6.3 vs 2.6 events per 100 patient-years, and acute kidney injury at 12.2 vs 6.7 events per 100 patient-years (both P<0.001). Harm outweighed benefit, and the trial was stopped early.
What matters most is that proteinuria had fallen further still. The surrogate improved; the patient did not. KDIGO 2024 writes this lesson directly into its recommendation: in patients with CKD, regardless of diabetes status, avoid combining an ACE inhibitor, an ARB, and a direct renin inhibitor in any combination (1B).
From then on, the principle settled into "maximum tolerated dose of a single agent." If you're going to add something, add a different pathway — not the same one.
Act Two: The Second Pathway Came From an Unexpected Place
The second pillar did not come from nephrology. It came from a cardiovascular safety trial of a diabetes drug. EMPA-REG OUTCOME (2015, 7,020 patients, median follow-up 3.1 years) tested empagliflozin in type 2 diabetes with established cardiovascular disease.
The result: 3-point MACE at 10.5% vs 12.1%, HR 0.86 (95.02% CI 0.74–0.99), P=0.04. But look at the components: myocardial infarction and stroke individually showed no significant difference. The difference came from cardiovascular death (3.7% vs 5.9%), heart failure hospitalisation (2.7% vs 4.1%), and all-cause mortality (5.7% vs 8.3%). This was not an atherosclerosis drug. It was a heart failure and kidney drug.
From there, trials with kidney outcomes as the primary target followed. CREDENCE (2019, 4,401 patients, all on background RAS inhibitor) showed canagliflozin reduced the primary composite with HR 0.70 (0.59–0.82) and end-stage renal disease with HR 0.68 (0.54–0.86). DAPA-CKD (2020, 4,304 patients, roughly a third without diabetes) showed the primary composite at 9.2% vs 14.5%, HR 0.61 (0.51–0.72), and even all-cause mortality at HR 0.69 (0.53–0.88). EMPA-KIDNEY (2023, 6,609 patients) showed 13.1% vs 16.9%, HR 0.72 (0.64–0.82). All three trials were stopped early for benefit.
The eligible population widened — from "diabetic nephropathy" to "CKD with albuminuria," and further to "CKD regardless of diabetes or albuminuria status." But widening the population meant enrolling a progressively lower-risk group. In EMPA-KIDNEY, neither the composite of heart failure hospitalisation/cardiovascular death nor all-cause mortality reached significance.
9 more sections, 7 figures, drawing on 16 trials, follow.
Read on — 14 days at ¥0 Log inThis page is educational commentary on the medical literature and is not a substitute for clinical judgement in an individual patient. The figures given are those reported in the original papers. Decisions on whether to treat, on dose and on targets must rest on the original papers and on the guidelines that apply where you practise.