Anemia of Dialysis-Dependent CKD: How Far to Chase the Hemoglobin
A single patient on hemodialysis, asking for real relief from real fatigue. The disagreement isn't whether his hemoglobin could go higher — it's how much higher is actually safe for a man with his cardiac history.
Walter G. has lived alone in the house he and his wife bought in 1991 since she died two years ago, and still keeps her chair at the kitchen table set the way she liked it. He drove long-haul freight for thirty-four years before his kidneys retired him instead of the other way around, and has been on in-center hemodialysis three days a week for four years, and the one thing he has asked his nephrologist for at nearly every visit since is more energy — enough to make it to his granddaughter's Saturday soccer games without needing the whole next day to recover. His current darbepoetin dose has held his hemoglobin at 9.2 for the past several months, and he says plainly that he still feels like he's moving through water; his dialysis nurses have also noticed him dozing through parts of his four-hour sessions, which he attributes to the same fatigue rather than anything about the treatment itself.
His iron studies are replete — ferritin 340, transferrin saturation 28 percent — his PTH is controlled, and there is no bleeding source or inflammatory process dragging the number down. This is a case where the dose genuinely could be raised and the hemoglobin genuinely would likely climb in response; the marrow limitation here is erythropoietin drive, not substrate. What complicates the request is not whether a higher dose would work, but what a higher target has actually been shown to cost in patients who look like him.
He has known coronary artery disease from a stent placed six years ago after an exertional chest pain workup, and the trials that tested higher hemoglobin targets against hard outcomes, not just symptoms, were not reassuring in exactly that population. The proposed mechanism isn't subtle: raising hemoglobin increases blood viscosity and platelet reactivity, plausibly tipping a borderline coronary lesion toward thrombosis in a way that never shows up on a symptom questionnaire until the event itself. That is the specific, mechanistic reason his stent history belongs in this decision rather than being treated as an unrelated fact sitting alongside it.
How far to chase the number
He isn't asking for something unreasonable — a hemoglobin of 9.2 is genuinely symptomatic, and I have no trouble believing raising it would make him feel better. The question I can't get past is what the Normal Hematocrit Trial actually found when it tested this directly in dialysis patients: randomized to a hematocrit target of 42% versus 30%, the higher-target arm was stopped early for excess mortality and nonfatal MI, concentrated in patients with existing cardiac disease — which is exactly the box he's in.
CHOIR found the same directional signal in non-dialysis CKD, and TREAT, in diabetic CKD patients targeting near-normal hemoglobin with darbepoetin specifically, doubled the risk of stroke. None of these trials targeted the specific number he's asking for, but all three point the same direction: pushing this drug toward a higher hemoglobin target in a patient with known coronary disease has cost more than it's given back, every time it's been tested.
The trial data are real and I'm not arguing to chase a normal hemoglobin. But it's worth being precise about what those trials actually tested: they randomized to a TARGET, generally near 13–14 g/dL, not to "any dose increase from 9.2." A modest increase toward 10–10.5, with a hard ceiling and close monitoring, is a materially different exposure than what produced the harm signal in Normal Hematocrit or TREAT.
His symptom burden is real and current guidelines don't ask us to leave a symptomatic patient at 9.2 untouched — they ask us to avoid normalizing the number, which is a narrower instruction than avoiding any increase at all.
Agreed: a modest darbepoetin increase targeting 10–10.5 g/dL, with monthly hemoglobin checks and an explicit instruction to hold or reduce the dose if the trajectory overshoots toward normalization.
Not agreed: how much of his fatigue this increase will actually resolve. The nephrologist was candid that a hemoglobin of 10.5 is unlikely to fully answer his stated goal, and that expectation was set with him directly rather than left implicit.