RAAS Inhibition on the Approach to Dialysis
Two patients closing in on end-stage kidney disease test the same old instinct — that a RAAS inhibitor should come off before the kidneys 'take the hit' of dialysis planning. Recent trial evidence reverses that instinct in general. Whether it resolves either patient's own case is a separate question.
Twelve years without a single tolerance issue is the track record now up for debate. R.T. is a 68-year-old man who retired two years ago from thirty-five years running a small hardware store, and now spends most mornings restoring an old rowboat in his garage. He has taken lisinopril for CKD and hypertension for twelve years without incident — no hyperkalemia, no symptomatic hypotension, blood pressure consistently near goal. His CKD, attributed to longstanding hypertensive nephrosclerosis, has progressed gradually to an eGFR of 16, and his access surgeon placed an arteriovenous fistula last month in anticipation of dialysis over the coming year. At his pre-dialysis visit, the covering nephrology fellow raised the familiar teaching: stop the ACE inhibitor now, since the kidney's own filtration work is about to be handed off, and there's no remaining benefit worth the drug's own hemodynamic burden on a gland already failing.
That teaching predates a real answer to the question it assumes. STOP-ACEi randomized 411 patients with advanced, actively progressing CKD — eGFR under 30, already losing more than 2 mL/min a year, already on a RAAS inhibitor for at least six months — to continue or to stop. Its primary endpoint was eGFR at three years, and the two arms finished within a point of each other, 12.6 against 13.3, with the difference not significant and running, if anything, the wrong way for stopping. What the trial did not do is demonstrate cardiovascular benefit from continuing: it recorded 108 cardiovascular events in the discontinuation arm against 88 in the continuation arm, but its authors say plainly it was never powered to answer that question, and a numerical gap in an underpowered secondary count is not a finding. The cardiovascular case for keeping R.T. on lisinopril rests where it always did, on the wider RAAS-inhibitor evidence base; what STOP-ACEi removes is the reason he was about to be asked to give it up. His own numbers place him inside that trial's population with nothing pulling the other way — twelve years without hyperkalemia, a potassium of 4.4, no orthostatic symptoms at 124/72.
Continue the lisinopril. STOP-ACEi enrolled a population that looks like R.T. specifically — advanced CKD, most already eGFR under 20 — and randomized continuation against the exact instinct being proposed here. Continuation didn't accelerate his need for dialysis — eGFR at three years came out within a point either way, and the ESKD numbers, if anything, favored staying on. I'd be careful not to overclaim the cardiovascular half: STOP-ACEi wasn't powered for it and didn't demonstrate it. The protection he still benefits from is the protection RAAS inhibitors have always had evidence for, and this trial simply removes the reason we were about to take it away from him. There's nothing in his own tolerance history that argues for stopping a drug the trial itself found safe to keep running.
I'll concede the point directly — I was reasoning from teaching that predates the trial that actually tested it, and R.T.'s own numbers don't give me a patient-specific reason to override what STOP-ACEi found. I'd have stopped it reflexively a year ago; I don't have a good argument to stop it today.
Agreed without real disagreement once the fellow's initial instinct was checked against the trial evidence directly: lisinopril continued unchanged through fistula maturation and into dialysis planning, reassessed only if a new tolerance issue emerges.
M.O., a 71-year-old woman, taught piano out of her home for over forty years and still keeps three students on her weekly schedule, though she has had to move her lessons to the ground floor since climbing stairs began leaving her lightheaded. Her CKD, also attributed to long-standing hypertension, has reached an eGFR of 15, and like R.T. she is approaching dialysis with an AVF already maturing. Unlike R.T., her course on losartan has not been clean, and the pattern matters more than either number alone: two hyperkalemia episodes in six months, at 5.9 and then 6.1, the second one arriving after patiromer had already been added for the first — which makes this binder-refractory rather than simply untreated, a different problem with a different answer. Her lightheadedness on the stairs has likewise been worked up rather than assumed, and a 22-point systolic drop on repeat measurement puts it well past the threshold where orthostatic hypotension stops being a plausible label and becomes a documented one.
STOP-ACEi's finding — that stopping buys nothing, so there is no reason to stop reflexively — was built from a population entered on the basis of declining eGFR, not on the basis of tolerating the drug badly. Patients experiencing recurrent binder-refractory hyperkalemia and confirmed symptomatic hypotension on the agent under discussion are not what its arms were populated with, and a trial that answers "is there a reason to stop in general" does not thereby answer "is there a reason to stop in her." M.O. sits at the edge of what that average can honestly be extrapolated to cover: an individual-level safety signal the aggregate outcome doesn't erase, and one that predates any question about her kidneys.
The overall message still holds: continue where possible. STOP-ACEi's finding that stopping accelerates nothing is a population finding that doesn't stop applying just because a patient is harder to manage — the harder cases are exactly where the drug's protection matters most, since she's carrying real cardiovascular risk into dialysis either way.
I'd stop it. Two binder-refractory hyperkalemia episodes and confirmed symptomatic orthostasis aren't background noise the trial's average absorbed — they're the specific harms a RAAS inhibitor is known to cause, actively occurring in front of us. STOP-ACEi's population wasn't dominated by patients like her; applying its average result to her specific, ongoing toxicity isn't supported by what the trial actually tested.
I take the cardiovascular-protection point seriously, but protection that keeps producing dangerous potassium levels isn't protection she can safely keep taking.
Neither full continuation nor full discontinuation is actually what STOP-ACEi tested — that trial compared continuing at existing dose against stopping entirely, not dose reduction. Halving her losartan preserves meaningful RAAS blockade and a real share of the cardiovascular benefit while easing both problems: less aldosterone suppression means less potassium retention pressure on an already-struggling binder regimen, and less vasodilation means less orthostatic burden. It's a genuine third option sitting outside what either arm of the trial actually answered for her.
Agreed: losartan reduced rather than stopped, patiromer continued, potassium and orthostatic symptoms rechecked within a week to confirm the reduced dose actually resolves what full-dose continuation could not.
Not agreed: whether a dose reduction that fails to control her potassium should be followed by full discontinuation or by further reduction — the hospitalist would stop entirely at that point; the nephrologist would try a lower dose first. That branch point was left explicit rather than pre-decided, contingent on the recheck.