His Kidneys Are Failing Faster Than His Skin Is Improving
A man whose severe atopic dermatitis is well controlled on upadacitinib develops progressive diabetic nephropathy, and his declining eGFR is now close enough to the drug's labeled renal dose-reduction threshold that the question isn't whether to act, but exactly when the number crosses the line that requires it.
Nathaniel P., 64, retired two years ago after four decades as a union electrician, and has spent much of his newly available time helping his daughter renovate a house across town — work that, until recently, his skin let him do without the constant cracking and bleeding that defined years of poorly controlled severe atopic dermatitis. Upadacitinib, started sixteen months ago after two biologics failed to control his disease, brought his EASI score from 31 to 4 within ten weeks and has held it there since — the longest sustained remission he's had as an adult.
Type 2 diabetes, present for over a decade and managed with metformin and, more recently, insulin, has produced progressive diabetic nephropathy that has accelerated over the past eighteen months: his eGFR has fallen from 71 to 42 across four sequential measurements, a clearly downward trajectory rather than normal variation around a stable baseline. Upadacitinib's label draws exactly one renal line, and his 42 sits on the permissive side of it: for atopic dermatitis the maximum recommended dose falls to 15mg once daily in severe renal impairment, defined as an eGFR of 15 to under 30, while for mild and moderate impairment — everything from 30 up — the label states that no dosage adjustment is needed. There is no graded middle zone, which means the label is not the thing leaving room for judgment here. What leaves room is that he is 64, and the same label already drops the recommended dose to 15mg once daily at 65, whatever his kidneys are doing that year.
He has noticed the decline himself in small, practical ways before any lab confirmed it — more swelling in his ankles by the end of a work day at his daughter's house, a persistent low-grade fatigue he'd initially blamed on age rather than kidney function. What worries him more than the numbers, when asked directly, is the memory of his own father's slide into dialysis a decade ago. The limitation nobody in the room can design around is that the label speaks in fixed thresholds and his kidneys are not standing at one: every figure it offers describes a patient measured once, and he is a patient measured four times and still falling.
A trend line versus a threshold
I'd reduce the upadacitinib dose now rather than wait for a future lab to formally cross a threshold. His trajectory over four consecutive measurements — 71, 61, 51, 42 — isn't noise around a stable baseline, it's a clearly progressive decline, and by the time a fifth measurement confirms he's below 30, we've been dosing at a level the pharmacokinetic data may not actually support for however long that decline continues.
I want to be precise about what the prescribing information actually says, because I think it's being misremembered in this room. There is no 30-to-59 judgment zone. The label says no dose adjustment is needed in mild or moderate renal impairment, and drops the atopic dermatitis dose to 15mg only in severe impairment — eGFR 15 to under 30. At 42 he is squarely inside the range the label leaves unadjusted, and the pharmacokinetic work behind that line found mild and moderate impairment had no clinically relevant effect on upadacitinib exposure at either dose. Reducing today substitutes a trajectory projection for the measured value that dosing data are anchored to.
I don't think the nephrologist's trend-based reasoning is wrong as a clinical instinct — it's a genuinely reasonable way to think about a moving target — I'm just flagging that it isn't what the labeled threshold itself is measuring, and conflating the two risks either under- or over-correcting relative to what the pharmacokinetic data actually support at his current, still-above-30 eGFR.
Whichever way this resolves, I'd want it resolved as precisely as possible rather than moved earlier out of caution, because a dose reduction isn't free for him. His EASI held at 4 for over a year on the current dose after two prior biologics failed outright — a real, demonstrated fragility to his disease control that a premature dose cut could disrupt for a renal threshold he hasn't actually reached yet. I'd support a closer monitoring interval — repeat eGFR at four weeks instead of the usual three months — so we catch the actual crossing quickly rather than either waiting too long or acting too early on a trend.
Agreed: continue upadacitinib at the current dose, shorten the renal monitoring interval to four weeks, and pre-commit to an immediate dose reduction the moment eGFR crosses below 30 rather than waiting for a scheduled visit to notice it.
Not fully settled: whether the accelerating rate of decline itself, rather than the absolute number, should trigger earlier action if the next measurement shows the trend steepening further. The nephrologist said she would push to act on trajectory alone if the next value dropped by more than the roughly 9-10 points per interval seen so far; the pharmacologist remained committed to waiting for the labeled threshold itself unless a nephrology-specific rationale for anticipatory action could be more precisely defined than 'the trend is concerning.'