Reviews/ACE Inhibitors and Angiotensin Receptor Blockers/ACE inhibitor and ARB withdrawal worsens heart failure outcomes
Hfref · ACEARB-HFREF-04 · v2
This claim asks whether stopping two common heart-failure drugs (ACE inhibitors and ARBs) while a patient is in hospital for heart failure makes them more likely to die. The largest directly relevant study [4] found that patients whose drug was stopped during the hospital stay died more often at 30 days and at one year than those kept on it, even after accounting for how sick they were. A separate trial-based analysis [7] pointed the same way: patients who kept or raised their dose during an acute admission survived better. Other heart-failure records [6][8] showed the same direction, though not during a hospital stay. The weakness is that none of this comes from an experiment that deliberately stopped the drug in some patients and not others, so it is hard to be sure the drug was stopped because patients were already dying rather than the stopping itself causing the harm.
The on-point registry [4] reports 30-day mortality of 8.8% when ACEi/ARB was discontinued during hospitalisation versus 3.5% when continued (HRadj 1.92; 95% CI 1.32-2.81), and 41.6% versus 28.2% at one year (HRadj 1.35; 95% CI 1.13-1.61), even after adjustment for severity of illness. The trial-derived PROTECT analysis [7] shows the inverse relationship: patients who maintained or increased their ACEi/ARB dose during acute HF hospitalisation had lower 180-day mortality (HR 0.47; 95% CI 0.29-0.77). A chronic-HF registry [6] found all RAASi discontinuations strongly associated with mortality, and an improved-EF cohort [8] found RASi/ARNi withdrawal associated with a 38% higher risk of CV mortality or HF hospitalisation. Every included source [4][5][6][7][8] is observational. The contested point: The division is whether the withdrawal-mortality association is causal or whether ACEi/ARB are discontinued precisely in patients who are already dying (hypotension, cardiogenic shock, worsening renal function), making withdrawal a marker rather than a cause. No randomised trial in the corpus tests withdrawal during HF hospitalisation, so the question turns entirely on how fully observational adjustment [4] can close that gap.
All 29 papers behind this claim with their DOIs and what each does to it, the panel’s arbitration, the evidence gap map and the open research questions. Free, and sent to your inbox.
We’ll also send you other claims from ACE Inhibitors and Angiotensin Receptor Blockers, one at a time, and tell you when a release or a pivotal trial changes a grade. Unsubscribe by replying.
One reviewer's only job is to attack the finding. The challenges below were weighed before the grade was set.
Confounding by indication is the central threat: ACEi/ARB are withdrawn precisely in patients with hypotension, cardiogenic shock or worsening renal function - those already at highest mortality risk. Every included source is observational; [4]'s adjustment 'for severity of illness' narrows but cannot exclude that withdrawal is a marker of dying patients rather than a cause of death.
No randomised trial in the corpus tests ACEi/ARB withdrawal during HF hospitalisation. The only withdrawal RCT [2] is outpatient recovered-DCM with a relapse endpoint and zero deaths; the only discontinuation RCT [15] is in COVID-19, not HF, and found no significant mortality difference. The verdict rests entirely on observational association.
Where it remains uncertain. The signal cannot be separated from confounding by indication, since drugs are withdrawn in the sickest patients, and no randomised trial in the corpus tests the question; one adjacent cohort [19] even found neutral mortality on discontinuation (HR 0.93; 0.86-1.01).
Aleatory
Discontinuation was uncommon (1.9% in [4]; 4.4% RASi/ARNi withdrawal in [8]), so event counts driving the hazard estimates are modest and subject to chance variation.
Epistemic
No trial in the corpus randomly withdrew ACEi/ARB during HF hospitalisation; [2] is recovered DCM with zero deaths and [15] is COVID-19, leaving causation unestablished.
Inferential
[6] is chronic-HF outpatients and [8] is an improved-EF outpatient re-registration, not the during-hospitalisation population the claim specifies.
The load-bearing papers from a corpus of 29, each classified by what it does to the claim. Every DOI is taken from the corpus the claim was graded on.
Primary direct support: large registry (N=16,052) matching the claim's population, intervention and outcome exactly; 30-day mortality HRadj 1.92 (1.32-2.81), 1-year HRadj 1.35 (1.13-1.61).
“Compared with continuation, withdrawal of ACEi/ARB during heart failure hospitalization is associated with higher rates of postdischarge mortality and readmission, even after adjustment for severity of illness.”
Panel auditInternal validity: medium · scope of inference: Hospitalised HFrEF (EF ≤40%) patients eligible for ACEi/ARB in US practice.Key limitations: observational confounding by indication · reason for discontinuation unmeasured
Independent large registry (N=9,222) corroborating the withdrawal-mortality direction in chronic HF; indirect to the hospitalisation setting.
“Importantly, all RAASi (ACEi, ARB, or MRA) discontinuations were strongly associated with mortality.”
Panel auditInternal validity: medium · scope of inference: Chronic HF outpatients (mostly HFrEF) across 31 countries, ambulatory setting.Key limitations: outpatient not hospitalisation setting · discontinuation as risk marker · ACEi/ARB pooled with MRA
Trial-derived support from within acute HF hospitalisation: maintaining/increasing ACEi/ARB is protective, the mirror image of withdrawal harm.
“Patients with incident hyperkalemia who were discharged with the same or increased dose of MRAs (HR: 0.52; 95% CI: 0.32 to 0.85) or angiotensin-converting enzyme (ACE) inhibitors/angiotensin receptor blockers (ARBs) (HR: 0.47; 95% CI: 0.29 to 0.77) had a lower 180-day mortality.”
Panel auditInternal validity: medium · scope of inference: Patients hospitalised for acute decompensated HF enrolled in PROTECT.Key limitations: observational dose-change analysis · ACEi/ARB partly pooled with MRA in framing
The abstract reports discontinuation frequencies and a pooled event rate across groups, with no mortality comparison between discontinuation and continuation.
Panel auditInternal validity: medium · scope of inference: New users of GDMT after a recent HF hospitalisation in Sweden, UK and US.Key limitations: descriptive only · no mortality contrast reported
4 of 29 papers shown. The full review lists all of them, with the panel’s arbitration →
Claims are graded at each quarterly release, and re-graded between releases only when a pivotal trial lands. Every change is a new version of the claim.
Strong, consistent, independently replicated observational evidence directly addressing the claim supports it, but the absence of any randomised trial and an unresolved confounding-by-indication threat hold confidence below the Established band.
Release 2026-q4 re-graded every claim from its corpus under protocol v0.19. An earlier release's grade is shown for the record; it came from an earlier protocol and is not a movement.
Assumptions that carry weight but are not themselves established.
The panel records the gap that would most move the claim, and the study that would close it.
No randomised trial of withdrawal during hospitalisation. Every supporting source is observational, so the causal effect of stopping ACEi/ARB during an HF admission is untested and confounding by indication cannot be excluded. The panel rates this the highest-importance open gap for the claim, and names the study design that would close it: A randomised trial comparing continuation versus protocolised withdrawal of ACEi/ARB during HF hospitalisation, with all-cause mortality as the endpoint and prespecified subgroups for hypotension and renal function.
Open question: Does deliberately withdrawing ACEi/ARB during HF hospitalisation increase mortality compared with continuation, independent of why the drug was stopped? Why it is still open: No randomised trial in the corpus tests withdrawal during hospitalisation; the only withdrawal RCT [2] had zero deaths and the only discontinuation RCT [15] was in COVID-19.
Priors is a sceptical first reader, not peer review, and not medical advice.
Release 2026-q4 · graded 21 Oct 2026 · protocol v0.19