Diuretic Resistance in Cardiorenal Syndrome: Reaching for a Third Nephron Segment
A single patient, four days into decompensated heart failure. The disagreement isn't about whether he needs more decongestion — it's about which second agent earns its risk in a kidney that was already running short before any of this started.
D.M., a 68-year-old man, has spent the past several years building furniture for whichever member of his family needed a table, a bed frame, or a set of shelves next; the current project — a dresser for a granddaughter due in the spring — has sat untouched in his garage since he twisted his knee on a scrap of lumber ten days ago. He had been managing the pain on ibuprofen from the medicine cabinet, several times a day, well past what the label recommended, until his daughter noticed his ankles disappearing into his shoes and brought him to the emergency department.
He has carried a diagnosis of heart failure with reduced ejection fraction (HFrEF) for three years, EF 30% on his last echocardiogram, and had been stable on a full four-drug regimen — sacubitril-valsartan, metoprolol succinate, spironolactone, and dapagliflozin — with no hospitalization since diagnosis. His kidney function has never been normal: an eGFR around 45 mL/min/1.73m², attributed to a decade of hypertension and type 2 diabetes rather than to the heart failure itself, tracked at clinic visits but never treated as urgent. Four days ago, twenty pounds up from his dry weight and short of breath lying flat, he was converted from his home oral torsemide to IV furosemide and diuresed briskly at first — nearly two liters net negative on the first hospital day alone. That response has not held. Over the past forty-eight hours, despite the furosemide dose climbing to 160 mg IV twice daily, his net output has flattened to under half a liter a day, his weight hasn't moved, and his creatinine has drifted from a baseline of 1.3 up to 1.6 — the pattern the team on rounds this morning is calling diuretic resistance riding on top of an already-reduced renal reserve, not a straightforward volume problem that simply needs a bigger dose of the same drug.
The mechanism is why simply pushing the same drug harder is running out of room. Furosemide works at the thick ascending limb by blocking the NKCC2 transporter; sustained loop-diuretic exposure drives compensatory sodium reabsorption downstream in the distal nephron, the same adaptive process that makes chronic loop-diuretic use progressively less effective at a fixed dose — a real physiologic phenomenon, not simply a matter of insufficient dosing. Two drugs approach that downstream adaptation differently: a thiazide-type agent blocks the NCC transporter directly in the distal convoluted tubule, the very segment doing the compensating, while acetazolamide works further upstream at the proximal tubule, inhibiting carbonic anhydrase and blocking bicarbonate-linked sodium reabsorption before it ever reaches the loop — increasing the sodium load the loop diuretic has to work with, rather than blocking a parallel escape route outright. Whether that mechanistic distinction actually buys a meaningfully different renal safety margin in a patient already down to an eGFR of 45, or whether it matters less in practice than the trial data suggest, is exactly what's under debate this morning.
At the bedside, day four
Residual congestion, not a creatinine of 1.6, is what predicts his readmission and his mortality — that's consistent across every large heart failure registry, and four days in he still has rales, a JVP of 12, and edema to the knees. Standing pat and just watching is not a neutral choice; it has its own cost. I want to add acetazolamide, 500 mg IV daily, on top of the current furosemide dose rather than raising furosemide any further. ADVOR randomized more than five hundred patients in a comparable decompensation and found a meaningfully higher rate of successful decongestion with early acetazolamide, more diuresis and natriuresis, and no worse mortality or heart failure readmission at three months. The trial's worsening-renal-function signal was real but transient and didn't track with worse outcomes — a rising creatinine that resolves is not the same finding as a rising creatinine that predicts a bad one, and ADVOR is direct evidence the two can come apart.
And to head off the obvious alternative: I don't think a continuous furosemide infusion buys us anything here. DOSE tested continuous infusion against twice-daily bolus dosing head-to-head and found no meaningful difference in symptom relief or renal outcomes at the same total dose — it would add infusion-pump logistics without addressing the actual resistance mechanism.
I'm not arguing residual congestion doesn't matter. I'm arguing this kidney doesn't have much room left, whatever the trial's average enrollee looked like. A baseline eGFR of 45 means he's already lost more than half his nephron mass before any of this started; a creatinine rise from 1.3 to 1.6 in someone starting from a normal baseline is a different event, physiologically, than the identical rise in someone already at CKD stage 3a — proportionally, he's spending a larger fraction of what he has left. ADVOR's own renal-safety data — worsening renal function in roughly forty percent of the acetazolamide arm versus under twenty percent on placebo — is not nothing, even though it washed out by three months. And there's a cleaner mismatch than any of that: ADVOR excluded patients taking an SGLT2 inhibitor outright. He's on dapagliflozin, which acts at the same proximal segment acetazolamide does. Whatever that trial measured, it did not measure adding acetazolamide on top of a drug already working upstream.
And ADVOR added acetazolamide up front, at randomization, in patients who hadn't yet been shown to be diuretic-resistant on their admission loop-diuretic dose. He has already failed four days of loop-diuretic escalation with a creatinine that's actively climbing — that's a different starting point than the trial tested, and I'm not willing to assume the same safety margin carries over to a kidney already trending the wrong direction before we've even added the second drug. My preference is to hold any second agent for now — no acetazolamide, no thiazide-type addition — recheck his creatinine and urine output in twenty-four hours at the current furosemide dose, and let his own trajectory tell us whether this is genuine diuretic resistance or an ordinary lag in a kidney that needed one more day to catch up. If he's still flat at that point, I'll reconsider.
I don't think either of you has to be wrong here, and I don't think the choice is simply "cautious" versus "aggressive" — it's about which second agent, if any, actually earns its risk in this specific patient. A thiazide-type addition — hydrochlorothiazide or metolazone — is the one option I'd take off the table outright. CLOROTIC showed it does exactly what it's supposed to: better weight loss, better diuretic response. But it came at a real, measured cost in renal function and electrolyte disturbance — and that trial's own population had a similarly reduced baseline eGFR, in the low forties, to his. Even there, the added natriuresis didn't translate into any difference in mortality or rehospitalization at ninety days. That's the worst combination — real downside, no proven upside beyond a surrogate, in a population not so different from him.
Acetazolamide is mechanistically a different proposition. It isn't competing with furosemide for the same downstream segment the way a thiazide is; it's increasing what reaches the loop in the first place, which is closer to actually reversing the adaptation driving the resistance rather than layering on a second blunt instrument at the same site. And the outcomes data — imperfect population match aside — are real: more successful decongestion, no worse mortality or readmission. I'd add it now, at the standard dose, specifically because it's the one addition with genuine trial support behind it. But I want the recheck interval tighter than either of you has proposed, and an explicit number for when we stop — not a vague "reassess tomorrow."
Agreed: hold furosemide at 160 mg IV twice daily rather than escalating further or converting to a continuous infusion, add acetazolamide 500 mg IV daily, and take the thiazide-type option off the table for this admission. Recheck creatinine, potassium, and net fluid balance at 24 and 48 hours rather than the usual 48-hour-only interval, given the baseline renal reserve already in question.
Not agreed, and the reason the monitoring plan carries an explicit branch point rather than a shared expectation:
Continue the current regimen and reassess daily — the nephrologist's threshold for real concern, not yet crossed.
Acetazolamide is held immediately, nephrology takes primary direction of further decongestion, and ultrafiltration returns to the table as a real option rather than a deferred one.
The cardiologist and the nephrologist left rounds with genuinely different expectations for where that number lands over the next two days — one reading his overall trajectory as a kidney that will tolerate one more push, the other reading the same numbers as a kidney already telling them to stop asking.