Treating Uremic Itch in a Frail Woman Who Has Chosen Against Dialysis
With the decision against dialysis already made and settled, the question left is a narrower one: how to actually treat disabling itch in a woman whose kidneys can no longer clear the drug with the best evidence for treating it.
Agnes W., an 88-year-old woman, raised four children largely on her own after her husband's early death and has lived independently in the same house for sixty years, though her daughter now stays most nights since Agnes's kidney function crossed into territory that made living entirely alone feel unwise to everyone involved. Her ESKD, eGFR now 9, was discussed at length with her, her daughter, and her nephrology team over several visits, and the decision against dialysis was made deliberately and is not in question today — Agnes was clear that she wanted her remaining time spent at home, not tethered to a schedule of treatments she didn't want. What has brought the team back together this visit is a narrower, still-real problem: pruritus so severe it keeps her awake most nights, scratching until her forearms show broken skin, distress her daughter says has become the dominant fact of Agnes's days.
The drug with the strongest evidence for uremic pruritus specifically is gabapentin — unlike antihistamines, which carry weak evidence for this particular symptom since uremic itch isn't primarily histamine-mediated. Gunal's randomized placebo-controlled crossover trial found real benefit, and later trials have reproduced it. But the population those trials studied is not Agnes's population, and the difference runs in the direction that matters: Gunal's patients were on hemodialysis, dosed at 300mg after each session, and the regimen works partly because dialysis itself removes the drug three times a week. Agnes has chosen not to have that. Her eGFR of 9 is the only clearance she has, gabapentin is renally cleared, and the schedule the evidence is built on assumes a route of elimination she doesn't possess — so accumulation risk here exceeds what those trials measured rather than matching it. Sedation, confusion, and falls land with particular force on an 88-year-old already frail enough that her family has reorganized their lives around keeping her safely at home.
Weighing a drug's accumulation risk against the cost of leaving her untreated
I'd start gabapentin, dosed for her kidney function. Gunal's randomized trial found real benefit for uremic pruritus specifically — antihistamines don't have that, since this isn't a histamine-mediated itch, whatever the reflexive first reach for a rash or itch tends to be. I'll grant the trials were done on dialysis, but the itch is the same itch and it is what's wrecking her nights. Agnes's goals are comfort, and genuine symptom relief is the actual clinical target here, not a secondary consideration.
That caveat isn't a footnote, it's the whole problem. Those patients cleared the drug three times a week on a machine; Agnes has declined the machine. At an eGFR of 9 with no dialytic clearance, gabapentin accumulates substantially even at doses already reduced for kidney disease, and the risk — sedation, confusion, falls — lands hard on a woman this frail, whose family has restructured their own lives specifically to keep her safe at home. I'd rather try a lower-risk option first, even with weaker evidence, given how costly an adverse event would be for her right now.
I'm not dismissing how real her distress is — I just think the asymmetry between a modest, uncertain benefit from a weaker option and a serious harm from the stronger one is worth taking seriously before we reach for the more effective drug.
Both of you are naming a real risk, and I don't think either should simply win. Standard "reduced ESKD dosing" tables aren't reduced enough for someone at her level of kidney failure and frailty — go well below even that, at the lowest plausible effective dose, and give her daughter explicit, specific red flags to watch for given how much time she already spends with Agnes. But I'd push back gently on treating caution as the automatically safer default: in a clearly comfort-focused context, leaving severe, sleep-disrupting distress untreated is also a real harm, not a neutral choice. Start low, watch closely, and treat this as titration toward relief rather than a decision between treating and not.
Agreed: gabapentin started at a dose well below standard ESKD-reduced tables, with Agnes's daughter given explicit sedation and confusion red flags to watch for and a direct line to call if any appear, and a follow-up check within a week rather than waiting for a routine visit.
Not resolved as a general rule: how far below standard dose-reduction tables a frail, very elderly ESKD patient's starting dose should go — handled here as an individualized judgment for Agnes specifically, not established as a new fixed protocol for conservative management patients broadly.