Anti-GBM Disease, Dialysis-Dependent at Presentation: How Hard to Treat a Kidney That May Not Recover
A single patient, dialysis-dependent at diagnosis with near-complete crescents on biopsy. The disagreement is whether treatment intensity should be set by the historic recovery odds at this severity, or by what the treatment is actually protecting against.
Harold J. spent ten days feeling generally unwell — fatigue he attributed to a bad flu season, decreased urine output he didn't think to mention until his wife noticed he'd stopped drinking his usual amount of coffee because it seemed pointless — before an emergency department visit found a creatinine of 8.9 and started him on dialysis the same day. The biopsy, obtained within forty-eight hours, showed 90% of glomeruli with crescents and the linear basement-membrane staining pattern that confirms anti-GBM disease rather than leaves it presumptive. His anti-GBM antibody titer came back strongly positive. He has no pulmonary hemorrhage — chest imaging is clear, no hemoptysis — which matters, because it changes what the team is actually treating for.
Anti-GBM disease's historic cohort data, dating back to Levy's foundational 2001 study, is honest about something uncomfortable: patients who present dialysis-dependent with this degree of crescentic involvement have very low rates of renal recovery, regardless of how aggressively they're treated. That's not a reason treatment doesn't matter — it's the reason how aggressively to treat a kidney with this little statistical chance of recovering becomes a real, separate question from whether to treat at all. Full-intensity therapy — daily plasma exchange alongside cyclophosphamide and pulse corticosteroids — carries genuine costs of its own: infection risk, bleeding and vascular-access complications from repeated PLEX, cumulative alkylator exposure. Whatever intensity is chosen, the antibody-suppression goal matters independent of the kidney: circulating anti-GBM antibody is what drives the disease's other genuinely life-threatening complication, pulmonary hemorrhage, which he doesn't have yet but which the same untreated antibody could still produce. His wife, sitting at the bedside through both conversations so far, has asked the same question twice in different words — whether any of this gets him off dialysis — and the honest answer the team keeps circling back to is that the antibody-suppression goal and the renal-recovery goal, close together in most kidney diseases, have genuinely pulled apart in his case.
On the ICU nephrology consult, day 2
Ninety percent crescents and dialysis-dependent at presentation is severe, I know the historic recovery odds are low — but they're not zero. Some patients at this severity do recover renal function, and I don't think we should foreclose that on his behalf. Full-intensity PLEX, cyclophosphamide, and pulse steroids, and we keep going as long as it takes.
I'd read the same historic data differently. At this severity, renal recovery is genuinely uncommon — most patients presenting dialysis-dependent with this degree of crescent involvement stay dialysis-dependent regardless of intensity. Plasma exchange isn't free; repeated exchanges carry real infection and access-complication risk. I'd treat with cyclophosphamide and steroids to suppress antibody production and protect against pulmonary hemorrhage, without committing to extended PLEX for a recovery chance this small.
'Some patients recover' is true, but it's also true at every severity level in that cohort — the question is whether the odds at HIS severity justify the added PLEX-specific risk, and the historic data actually argue they don't, clearly, at this level.
I don't think this has to be a choice between accepting a small chance indefinitely or writing it off entirely. Both of you are treating intensity and duration as the same decision — I'd separate them. Start full-intensity: daily PLEX plus cyclophosphamide and pulse steroids, giving him the real, if reduced, chance the historic cohorts describe.
But set an explicit checkpoint — two to three weeks — and de-escalate PLEX specifically if there's no recovery signal by then, even while continuing antibody-suppressing therapy for the pulmonary-hemorrhage protection that matters regardless of his renal trajectory. That way we're not committing him to open-ended PLEX risk once the realistic recovery window has passed, and we're not denying him the chance up front either.
Agreed: full-intensity plasma exchange, cyclophosphamide, and pulse-then-taper corticosteroids, with an explicit two-to-three-week checkpoint for de-escalating PLEX specifically if there's no recovery signal. Antibody-suppressing therapy (cyclophosphamide, steroids) continues regardless of that checkpoint's outcome.
Not fully agreed: whether the checkpoint should be two weeks or three — the reduced-intensity nephrologist preferred the earlier point, the full-intensity nephrologist the later one, and the difference was left as a judgment call for the treating team closer to the actual date rather than fixed now.