Skip to content
Accelerated desensitization enables HLA-incompatible deceased donor kidney transplant in a highly sensitized pediatric recipient
All organs

CHECKPOINT-INHIBITOR TOXICITY

Kidney immune-related adverse events

Acute tubulointerstitial nephritis dominates; glomerular lesions and electrolyte disturbances are increasingly reported.

The atlas covers kidney irAEs in depth elsewhere; this card is the organ-indexed view of the same content. Educational reference, not medical advice.

Anchored on the ASCO and ESMO chapters for this organ. Every source is named beside the position it supports, and labeled with what kind of source it is.

At a glance

ICI-associated acute tubulointerstitial nephritis

Also called: ICI-AKI · ICI-AIN · checkpoint inhibitor nephritis · immune checkpoint nephritis

The dominant kidney irAE: steroid-responsive interstitial nephritis presenting weeks to months into therapy, often with a concomitant AIN-inducing drug on board.

UncommonSeriousonset: month 3 onward

Reported frequency

  • 2-5% — any-grade ICI-AKI incidence

    of patients treated with ICIs · ICI-treated oncology patients, as synthesized in a nephrology review — a summary assertion with no denominator, which is why it is carried at review weight beside the pooled figure below

    narrative review · 36168055

  • 1.4% (95% CI 1.0-2.1%), incidence rate 4.3 per 100 patient-years (95% CI 2.3-6.3) — pooled one-year incidence of ICI-associated AKI

    of 10,726 participants across 16 studies · Systematic review and meta-analysis. Materially lower than the 2-5% carried above, which is a review's summary assertion with no denominator — both are shown rather than reconciled

    pooled meta-analysis · 41026215

Onset

Median 14 weeks from ICI initiation, IQR 6-37 weeks — the interquartile range alone spans early-weeks to late-months presentation, so a late AKI does not argue against the diagnosis.

Presentation

  • · Subnephrotic proteinuria in most patients
  • · Pyuria in approximately half
  • · Rise in serum creatinine, frequently while an AIN-inducing co-medication is in use

Differential

  • · Alternative causes of AKI — prerenal, obstructive, and other nephrotoxic exposures — which every anchor guideline requires be excluded first
  • · AIN attributable to a concomitant drug rather than the ICI: 69% of patients in the multicenter cohort were concurrently receiving a potential tubulointerstitial-nephritis-causing medication

Work-up

  • Serum creatinine trend against baseline · All patients

    Establishes the CTCAE grade that indexes every management step below.

  • Urinalysis with microscopy and urine protein-to-creatinine ratio · All patients

    Subnephrotic proteinuria and pyuria are the reported urinary phenotype.

    31896554

  • Review of concomitant AIN-inducing medications (PPIs, NSAIDs) · All patients

    Stopping them is part of the intervention, not only the differential. In the multicentre cohort 69% were concurrently receiving a potential tubulointerstitial-nephritis-causing medication.

    31896554

  • Kidney biopsy · Grade 2+

    Confirms interstitial nephritis versus an alternative lesion and informs treatment. The cited source scopes this to moderate-to-severe disease in words, not to a numeric grade — the grade threshold below is this atlas's reading of that phrase, not a quoted one.

    36168055

Grade ladder

Grade 1
Continue the ICIOutpatient

Grade 1 — creatinine 1.5-2.0x above baseline (CTCAE v5.0, Acute kidney injury).

No corticosteroid at this grade

ESMO permits continuing the ICI with monitoring at stage 1 AKI; the intervention at this rung is exclusion of alternative causes and withdrawal of concomitant nephrotoxins, not immunosuppression.

  • · Exclude alternative etiologies of AKI
  • · Stop concomitant nephrotoxins (PPIs, NSAIDs) where possible

Escalate when: Creatinine rising to grade 2, or failure to improve after nephrotoxin withdrawal.

Grade 2
Hold the ICIOutpatient

Grade 2 — creatinine >2.0-3.0x above baseline (CTCAE v5.0, Acute kidney injury).

Corticosteroid 0.5–1 mg/kg/day (oral)

Taper over 4–6 weeks or longer, starting once: Creatinine improving toward baseline.

No step size or interval is carried: the chapter cited does not specify a taper cadence, and one is not invented here.

  • · Exclude alternative causes before attributing the event to the ICI
  • · Stop concomitant AIN-inducing medications

Sources differ here

Grade definition

  • MD Anderson Cancer Center, Department of Clinical Effectiveness V3 (institutional, approved 06/2026) (institutional protocol) — Grade 2 (G2): creatinine > 1.5-3 times above the baseline. A patient at 1.8x baseline is grade 1 by CTCAE and grade 2 here — and grade 2 is where the ICI is held and corticosteroids start.
  • CTCAE v5.0 (standard grading vocabulary) — Grade 2 — creatinine >2.0-3.0x above baseline. CTCAE v5.0 'Acute kidney injury', which is the vocabulary this entry's ladder is written in and which no cited guideline chapter reproduces.

Taper

  • ASCO 2021 (society guideline) — tapered over 4-6 weeks once creatinine improves
  • MD Anderson Cancer Center, Department of Clinical Effectiveness V3 (institutional, approved 06/2026) (institutional protocol) — Once SCr starts to improve, start taper over 2-4 weeks.

Escalate when: No creatinine improvement on steroids, or progression to grade 3.

Grade 3
Hold — consider permanent discontinuationConsider admission

Grade 3 — creatinine >3.0x baseline or >4.0 mg/dL (CTCAE v5.0, Acute kidney injury).

Corticosteroid 1–2 mg/kg/day (oral)

Taper over 4–6 weeks or longer, starting once: Creatinine improving toward baseline.

No step size or interval is carried: the chapter cited does not specify a taper cadence, and one is not invented here.

Second line

  • Infliximab, added to glucocorticoids — Steroid-refractory or relapsed disease. In a 55-patient comparative series the advantage was tied to timing — infliximab started within 30 days of steroids shortened time to renal response — and it did not translate into a progression-free survival difference. An earlier 10-patient series had two patients with no improvement at all.Contraindicated in hepatic impairment, heart failure, and tuberculosis exposure; screen for latent TB and hepatitis B before the first dose.
  • Mycophenolate mofetil — Glucocorticoid-refractory disease, and — in the one reported case where steroids were contraindicated by concurrent ICI-induced diabetes — as first-line therapy instead of them. The evidence here is two single case reports, not a series: it establishes that the drug has been used successfully in this indication, and nothing about how often it works.
  • · Nephrology involvement
  • · Stop concomitant AIN-inducing medications

Escalate when: Steroid-refractory disease or progression to dialysis requirement.

Grade 4
Discontinue permanentlyAdmit

Grade 4 — life-threatening renal impairment, including dialysis requirement. The anchor guideline treats grade 3-4 as one dosing band but separates the discontinuation decision, which is why this rung repeats the grade 3 dose and diverges on ICI action.

Corticosteroid 1–2 mg/kg/day (oral)

Taper over 4–6 weeks or longer, starting once: Creatinine improving toward baseline.

No step size or interval is carried: the chapter cited does not specify a taper cadence, and one is not invented here.

Second line

  • Infliximab, added to glucocorticoids — Steroid-refractory or relapsed disease. In a 55-patient comparative series the advantage was tied to timing — infliximab started within 30 days of steroids shortened time to renal response — and it did not translate into a progression-free survival difference. An earlier 10-patient series had two patients with no improvement at all.Contraindicated in hepatic impairment, heart failure, and tuberculosis exposure; screen for latent TB and hepatitis B before the first dose.
  • Mycophenolate mofetil — Glucocorticoid-refractory disease, and — in the one reported case where steroids were contraindicated by concurrent ICI-induced diabetes — as first-line therapy instead of them. The evidence here is two single case reports, not a series: it establishes that the drug has been used successfully in this indication, and nothing about how often it works.
  • · Nephrology involvement
  • · Renal replacement therapy where indicated
  • · Pooled analysis found no additional benefit from combining intravenous with oral steroids (OR 0.928, P = 0.863), which is why no pulse regimen is carried on this rung

Escalate when: Already the top rung — failure to recover carries the prognostic weight below.

Rechallenge

Individualize the decision

Roughly a quarter of rechallenged patients in the multicenter cohort developed recurrent ICI-associated AKI, and a single-center series records both uneventful rechallenges and one recurrence with a fatal extra-renal event. The decision turns on recovery, severity and whether an effective alternative therapy exists.

  • 23% — recurrent ICI-associated AKI after rechallenge

    of patients rechallenged with an ICI after ICI-associated AKI · Multicenter cohort of 138 patients with ICI-associated AKI, 22% of whom were rechallenged

    cohort · 31896554

  • · Kidney function recovered, and the degree of recovery explicitly weighed
  • · Co-management with nephrology
  • · Concomitant AIN-inducing drugs stopped and kept off

Deliberately not carried

  • Grade 3 or higher share — No source cited here reports the grade 3+ share of ICI-AKI as a separate figure. Omitted rather than derived from the recovery distribution, which measures something else.
  • Mortality — The multicenter cohort reports that failure to achieve kidney recovery was independently associated with higher mortality, but states no case-fatality rate. A rate would have to be invented, so none is carried.
  • Difference between ICI classes — Combination ICI therapy was reported as an independent risk factor for developing ICI-AKI, but that is a risk-factor association rather than a measured class-stratified incidence, so it is not entered as a skew framing.
  • Second-line therapy — mycophenolate dose — Mycophenolate is carried as an agent on the evidence of two single case reports, but NO dose is encoded. The only schedule available is an institutional protocol's (500 mg every 12 hours titrated over 2 weeks to 1 g every 12 hours, maximum 3 months), and an institutional protocol may not set a number alone. There is no society position on a renal mycophenolate dose to place beside it, so it cannot even be carried as a divergence — a divergence needs two voices. This row previously asserted that no second-line agent at all was supported here, which became false once the infliximab and mycophenolate evidence was verified.
  • Grade definitions on the ladder — The bands are CTCAE v5.0 'Acute kidney injury' as named, not a quotation from any cited guideline — no anchor chapter carried here reproduces the CTCAE table, and CTCAE is not a PubMed record. Named explicitly because the previous strings were unsourced prose: grade 1 read 'creatinine rise above baseline without meeting grade 2 criteria', which graded a 1.2x rise as grade 1 and routed it to a rung that continues the ICI, when CTCAE puts 1.2x below grade 1 entirely.
  • Taper step size and interval — The ASCO recommendation gives a total window and a response marker — tapered over 4-6 weeks once creatinine improves — and no step size or interval. Earlier drafts carried 25% every 1-2 weeks, which came from the beta taper planner's illustrative 'e.g. by ~25% every 1-2 weeks' rather than from the cited recommendation. Nulled rather than re-attributed: no anchor guideline states a cadence for any organ.

Sources

  • Gupta S, et al. (2026) Immune Checkpoint Inhibitor-Associated Acute Kidney Injury: A Report from the 34th Acute Disease Quality Initiative (ADQI) Consensus Conference.Cited for: Consensus framing of ICI-AKI: the share of AKI attributable to ICIs, acute tubulointerstitial nephritis as the dominant biopsy lesion, biopsy as the diagnostic gold standard with no reliable distinguishing clinical feature, biomarkers not yet routine, earlier glucocorticoid initiation associated with higher kidney recovery, and the recurrence fraction on rechallenge.Supporting text: the PubMed abstract (checkable at the link above)
  • Cortazar FB, et al. (2020) Clinical Features and Outcomes of Immune Checkpoint Inhibitor-Associated AKI: A Multicenter Study.Cited for: Onset window, dominant histology, recovery distribution and rechallenge recurrence in ICI-associated AKI.Supporting text: the PubMed abstract (checkable at the link above)
  • Sprangers B, Leaf DE, Porta C, Soler MJ, Perazella MA (2022) Diagnosis and management of immune checkpoint inhibitor-associated acute kidney injury.Cited for: Incidence of ICI-AKI among ICI-treated patients, steroid responsiveness, and the biopsy threshold.Supporting text: the PubMed abstract (checkable at the link above)
  • Manohar S, et al. (2020) Acute Interstitial Nephritis and Checkpoint Inhibitor Therapy: Single Center Experience of Management and Drug Rechallenge.Cited for: Stopping concomitant AIN-inducing drugs alongside the ICI hold, and small-series rechallenge outcomes.Supporting text: the PubMed abstract (checkable at the link above)
  • Rao Ullur A, Côté G, Pelletier K, Kitchlu A (2023) Immunotherapy in oncology and the kidneys: a clinical review of the evaluation and management of kidney immune-related adverse events.Cited for: AIN as the most frequent kidney irAE, with glomerular and electrolyte phenotypes increasingly reported.Supporting text: the PubMed abstract (checkable at the link above)
  • Schneider BJ, et al. (2021) Management of Immune-Related Adverse Events in Patients Treated With Immune Checkpoint Inhibitor Therapy: ASCO Guideline UpdateParaphrase, not a quotation — this atlas’s condensation of the recommendationCited for: Grade-indexed hold, steroid dose and taper for ICI-related nephritis/AKI. Cited for the kidney only — this recommendation is scoped to nephritis/AKI and is not carried onto any other organ's card.Supporting text: this atlas's condensation
  • Haanen J, et al. (2022) Management of toxicities from immunotherapy: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-upParaphrase, not a quotation — this atlas’s condensation of the recommendationCited for: Excluding alternative causes, stopping concomitant nephrotoxins, and continuing the ICI at stage 1 AKI.Supporting text: this atlas's condensation
  • Herrmann SM, Abudayyeh A, Gupta S, Gudsoorkar P, Klomjit N, Motwani SS, Karam S, et al. (2025) Diagnosis and management of immune checkpoint inhibitor-associated nephrotoxicity: a position statement from the American Society of Onco-nephrology.Cited for: The existence and scope of a subspecialty society consensus statement on ICI nephrotoxicity. Cited for scope only — the per-claim practice points are in the full text, which is not reachable here.Supporting text: the PubMed abstract (checkable at the link above)
  • Youssef N, Habeeb M, Chen H, Alhamal Z, Mamlouk O, Lin H, Page V, et al. (2025) Infliximab for Grade III or IV Immune Checkpoint Inhibitor Nephritis Clinical and Translational Evidence.Cited for: Infliximab added to glucocorticoids in stage III-IV ICI-AIN: comparative design, the timing that mattered, and the limits of what was shown.Supporting text: the PubMed abstract (checkable at the link above)
  • Lin JS, Mamlouk O, Selamet U, Tchakarov A, Glass WF, Sheth RA, Layman RM, et al. (2021) Infliximab for the treatment of patients with checkpoint inhibitor-associated acute tubular interstitial nephritis.Cited for: The earlier, smaller series behind infliximab in CPI-ATIN, including how many patients did not respond.Supporting text: the PubMed abstract (checkable at the link above)
  • Moku P, Bakow B, Muthiah A, Birkenbach M, Austin M (2023) Steroid Refractory Immune Checkpoint Induced Acute Interstitial Nephritis Salvaged by Mycophenolate Mofetil.Cited for: Mycophenolate as salvage in glucocorticoid-refractory ICI-associated acute interstitial nephritis. A single case.Supporting text: the PubMed abstract (checkable at the link above)
  • Jessel S, Austin M, Kluger HM (2021) Mycophenolate as Primary Treatment for Immune Checkpoint Inhibitor Induced Acute Kidney Injury in a Patient with Concurrent Immunotherapy-Associated Diabetes: A Case Report.Cited for: Mycophenolate as FIRST-LINE therapy for ICI-associated AKI in the specific case where corticosteroids are contraindicated. A single case.Supporting text: the PubMed abstract (checkable at the link above)
  • Ho CW, Kang NW, Yeh TH, Chuang MH, Tsai WW, Wang HY, Wu VC, Pan HC, Chen JY (2025) Immune checkpoint inhibitors-associated acute kidney injury: a systematic review and meta-analysis of incidence, kidney recovery, and recurrent risk.Cited for: Pooled incidence of ICI-AKI with a stated denominator, the benefit of corticosteroids on kidney recovery, the absence of added benefit from combining intravenous with oral steroids, and pooled recurrence after rechallenge.Supporting text: the PubMed abstract (checkable at the link above)

ICI-associated kidney allograft rejection

Also called: ICI-induced acute allograft rejection · acute rejection in kidney transplant recipients on checkpoint inhibitors · checkpoint inhibitor allograft rejection

Alloimmune rejection of a kidney allograft precipitated by checkpoint blockade. Every cohort here that reports a time to rejection reports it in days rather than months, and graft loss is a frequent end point in the datasets that count it — though the most recent kidney cohort reports less of it than earlier ones and calls the therapy viable. CTCAE has no term for this lesion, and no society recommendation this atlas holds grades it or states a corticosteroid dose for it.

Frequency unclearCriticalonset: weeks 1–6

Reported frequency

  • 42% (29 of 69) — acute rejection after ICI initiation in kidney transplant recipients

    of 69 kidney transplant recipients receiving ICIs between January 2010 and May 2020 · Multicenter retrospective cohort, kidney transplants only — the largest kidney-specific dataset carried here. The abstract also reports a much lower acute rejection rate in a non-ICI group of kidney transplant recipients, but that group was assembled stage-matched for the study's EFFICACY comparison rather than matched for rejection risk, which is why this card shows the contrast as the source prints it (see the divergence below) and does not convert it into a ratio.

    cohort · 33359528

  • 19 of the 29 who rejected — allograft loss among the kidney transplant recipients who rejected

    of the 29 of 69 kidney transplant recipients in this cohort who developed acute rejection after ICI initiation · Same cohort as the row above. Carried as its own row because it is the figure behind this card's statement that graft loss is a frequent end point — a claim that must not live in a prose string with no denominator. It is conditional on rejection, unlike the Legris and Saleem graft-loss rows below, which are shares of a whole cohort.

    cohort · 33359528

  • 26.5% — biopsy-proven acute rejection after ICI initiation in kidney transplant recipients

    of 34 kidney transplant recipients analysed across 6 transplant centres in France, 2015-2024 · Retrospective multicentre cohort, kidney transplants only, and deliberately the recent era. Both a stricter numerator than the first row (biopsy-proven only) and a later one, so the gap between the two figures is not purely temporal and neither source lets you say how much each difference contributed.

    cohort · 40755508

  • 15% — graft loss in kidney transplant recipients treated with ICIs, recent era

    of 34 kidney transplant recipients analysed across 6 transplant centres in France, 2015-2024 · Same cohort as the row above, and the finding on this card that runs hardest against its own headline: the authors present this as a LOWER incidence of graft loss than previously reported rates and conclude that ICI therapy is a viable option for these recipients. It is a share of the whole analysed cohort, not of those who rejected — the abstract reports its outcome breakdown across all analysed patients — so it is not comparable with the 19-of-29 row above.

    cohort · 40755508

  • 36.2% (95% CI 30.7%-41.7%) — acute rejection at 1 year after ICI initiation

    of 343 solid-organ transplant recipients pooled from 128 studies · Individual-participant-data meta-analysis of SOLID-ORGAN transplant recipients, of whom 70.9% were kidney recipients — it is not a kidney-only figure. Time-indexed to 1 year, unlike the crude cohort proportions beside it, so it is not directly comparable with them either.

    pooled meta-analysis · 40545616

  • 18.4% (95% CI 13.7%-23.1%) — graft loss at 1 year after ICI initiation

    of 343 solid-organ transplant recipients pooled from 128 studies · Same analysis and same solid-organ-wide population as the row above. A share of the whole pooled cohort at one year, not conditional on rejection.

    pooled meta-analysis · 40545616

  • 11 of 23 — allograft rejection, renal stratum of a solid-organ transplant series

    of 23 renal transplant recipients within a 39-patient institutional series and systematic review · MD Anderson institutional experience plus a systematic review of five databases; the abstract's results never separate the two, so nothing on this card attributes a result to one half. The kidney stratum is printed with its own denominator, which is why it is carried here rather than the 41% all-organ figure.

    narrative review · 30992053

  • 39.8% — allograft rejection in published cases of ICI use in solid-organ transplant recipients

    of 83 published cases of ICI use for cancer in solid-organ transplant recipients · Systematic review of published cases, all organs — not kidney-only, and case-report publication favours the eventful.

    narrative review · 32027461

  • 71.0% — progression to end-stage organ failure after allograft rejection

    of the rejection episodes within those 83 published cases — the abstract attaches the share to the rejections ('leading to end-stage organ failure in 71.0% of cases') and prints no count for them · Same all-organ published-case review. The share is carried as the source states it and is deliberately NOT converted into a number of patients, because the abstract gives no count to convert.

    narrative review · 32027461

  • 0 patients — the trial's primary endpoint — irretrievable allograft rejection with no evidence of tumour response

    of 17 kidney transplant recipients treated with nivolumab with baseline immunosuppression left unchanged · Multicentre single-arm phase 1 trial. The endpoint is a COMPOSITE and this is not a statement that no rejection occurred. Two facts travel with it: the trial stopped early because of COVID-19 restrictions, so 17 is the analysed population rather than a planned one, and increased creatinine was among the commonest grade 3-4 adverse events, in three (18%) of the 17.

    trial safety data · 35809595

  • 1 patient on nivolumab + tacrolimus + prednisone, and 2 more among the 6 who then received ipilimumab + nivolumab; the abstract refers to 3 patients with TRAL — treatment-related allograft loss (TRAL) under standardized tacrolimus + prednisone

    of 8 evaluable kidney transplant recipients with advanced cutaneous cancers · Prospective trial in which immunosuppression was standardized rather than left at baseline — tacrolimus trough 2-5 ng/mL plus prednisone 5 mg daily — and whose stated conclusion is that this regimen provides insufficient allograft protection AND compromises immune-mediated tumour regression. The counterweight to the trial above: leaving immunosuppression alone and standardizing it are different interventions with opposite conclusions, and both are shown.

    trial safety data · 38252910

  • 40%-50% versus 0%-12% — allograft rejection risk in kidney transplant recipients, early post-marketing data versus recent prospective trials

    of kidney transplant recipients treated with ICIs; the review states no denominator for either band · A narrative review's summary bands, carried at review weight for the same reason this atlas's AIN card carries an undenominated 2-5% beside a pooled figure: the source states the numbers and no denominator. The review attributes the difference to modification of maintenance immunosuppression before the ICI is started. Its abstract does not name the two prospective trials, so this card does not claim which they are.

    narrative review · 39479217

Onset

Early, and early is what separates this lesion from the interstitial nephritis on this atlas's other renal card. Three of the datasets cited here report a median and all three report it in DAYS: 24 days from ICI initiation to rejection in the 69-patient kidney cohort, 52 days in the 34-patient French kidney cohort, and 21 days (95% CI 19.3-22.8) in the 39-patient institutional series and systematic review. Two of those three fall inside this card's 1-6 week window; the French cohort's 52 days does not, so the bucket is the modal value rather than a consensus. The other datasets cited here report no median at all. The two patients in the educational case report presented at 6 days and at 2 weeks. No cited source reports an interquartile or full range, so none is shown.

Presentation

  • · A rise in serum creatinine in a kidney transplant recipient, typically within weeks of the first ICI dose — earlier than the interstitial nephritis on this atlas's other renal card, whose median is months into therapy.
  • · Usually against a background of maintenance immunosuppression that was reduced or stopped before the ICI: both patients in the educational case report were diagnosed with acute rejection in that setting, and the one trial built around this question left baseline immunosuppression unchanged for exactly that reason.
  • · A biopsy finding rather than a clinical syndrome. In the contemporary French cohort every rejection episode involved T cells, with additional humoral features in a minority, and the review held here states that allograft biopsy is required to confirm the diagnosis.
  • · No distinguishing symptom is reported in any source cited here — the presentation is graft dysfunction, found on bloods.

Differential

  • · ICI-associated acute tubulointerstitial nephritis, the native-kidney lesion this atlas's other renal card covers. It is a different lesion with a different treatment, it presents months rather than weeks into therapy, and only an allograft biopsy separates the two here.
  • · Rejection driven by the pre-ICI reduction in maintenance immunosuppression rather than by the checkpoint inhibitor itself. The two are routinely confounded — the phase 1 trial that left baseline immunosuppression unchanged exists because of it — and no retrospective cohort cited here separates them.
  • · Rejection that would have occurred anyway. The largest kidney cohort here records acute rejection in a non-ICI group of kidney transplant recipients too, at a much lower rate — though that group was assembled stage-matched for the efficacy comparison rather than matched for rejection risk, so it bounds the question rather than settling it.

Work-up

  • Serum creatinine trend against baseline in the allograft · All patients

    The presenting abnormality, and what reaches most of these cohorts' case definitions. It is not a grading instrument here: no source cited on this card indexes any management step to a creatinine band.

  • Allograft biopsy · All patients

    The review held here states that allograft biopsy is required to confirm the diagnosis. It is also what separates rejection from the other causes of graft dysfunction and what characterizes it: of the 12 transplant biopsies in the 39-patient institutional series and systematic review, nine (75%) showed acute rejection and five of those were T cell-mediated, and in the 34-patient French cohort every rejection episode involved T cells with humoral features in one third.

    39479217, 30992053, 40755508

  • Donor-derived cell-free DNA, measured serially · All patients

    Investigational, and the evidence here is three patients. In the prospective kidney-transplant trial dd-cfDNA was measured approximately once every 2 weeks as a potential predictor of rejection, and in 2 of the 3 patients with treatment-related allograft loss it rose 10 and 15 days before the serum creatinine did. The review held here puts non-invasive biomarkers no higher than a possible role in risk-stratifying before ICI start and in monitoring for rejection, and states that allograft biopsy is still what confirms the diagnosis.

    38252910, 39479217

  • Review of the maintenance immunosuppression regimen, including any reduction made before the ICI · All patients

    The only modifiable factor either large dataset identifies. In the 69-patient kidney cohort, mTOR inhibitor use (odds ratio 0.26, 95% CI 0.09-0.72) and triple-agent immunosuppression (0.67, 0.48-0.92) were associated with lower rejection risk; in the 343-recipient individual-participant meta-analysis, maintenance with steroids and mTOR inhibitors carried a lower risk than other immunosuppressive agents (hazard ratio 0.30, 95% CI 0.14-0.63). Whether the regimen was reduced before the ICI is also the confounder that makes every retrospective rate on this card hard to read.

    33359528, 40545616, 35809595

  • Donor-specific HLA antibody testing before the first dose · All patients

    Risk stratification before exposure rather than diagnosis after it. The one trial that left maintenance immunosuppression unchanged enrolled only recipients with no or low concentrations of donor-specific HLA antibodies, so its result does not extend to a sensitized recipient. The review held here likewise places non-invasive biomarkers at the pre-ICI risk-stratification step.

    35809595, 39479217

  • Joint transplant-nephrology and oncology review before starting the ICI · All patients

    The trade-off is the finding in every source here. The systematic review frames it explicitly as the difficult tradeoff facing oncologists and transplant specialists managing transplant recipients with cancer, and the case report's teaching points turn on patient-centred decision-making about that balance.

    32027461, 39444717

Sources differ here

Reported frequency

  • Murakami 2021 (multicentre kidney cohort) (primary evidence) — Following ICI treatment, 29 out of 69 (42%) patients developed acute rejection, 19 of whom lost their allograft, compared with an acute rejection rate of 5.4% in the non-ICI cohort.Kidney-only, and the largest such cohort carried here. The non-ICI rate in the same sentence comes from a comparator the abstract assembled stage-matched for the study's efficacy comparison — 23 recipients with squamous cell carcinoma and 14 with melanoma — rather than matched for rejection risk, so this card shows the sentence as the source prints it and does not convert the contrast into a ratio.
  • Legris 2025 (French multicentre kidney cohort) (primary evidence) — Biopsy-proven AR occurred in 26.5% of patients, at a median time of 52 d after ICI start.A stricter numerator than the cohort above — biopsy-proven only — and a later era, 2015-2024. Both differences push the rate down, and neither source lets you say how much each contributed.
  • Legris 2025 (French multicentre kidney cohort) (primary evidence) — Our study suggests that ICI therapy is a viable option for KTRs with poor-prognosis cancers, demonstrating a 38% tumor response rate and a lower incidence of graft loss (15%) compared with previously reported rates.The most recent kidney-only cohort's own conclusion, carried here because it is the position that runs hardest against this card's headline and against the older series beside it. Graft loss here is a share of the whole analysed cohort, not of those who rejected, so it is not the same quantity as the 19-of-29 figure in the incidence table.
  • Saleem 2025 (individual-participant-data meta-analysis) (primary evidence) — Acute rejection occurred in 36.2% (95% CI, 30.7%-41.7%) at 1 year, and 18.4% (95% CI, 13.7%-23.1%) experienced graft loss at 1 year.Solid-organ-wide, 70.9% kidney, and time-indexed to one year — so it is not directly comparable with the crude cohort proportions on either side of it, and it is not a kidney-only figure.
  • Carroll 2022 (phase 1 trial, baseline immunosuppression unchanged) (primary evidence) — No patients had irretrievable allograft rejection without evidence of tumour response.This is the trial's COMPOSITE primary endpoint — irretrievable rejection AND no tumour response — not a statement that no rejection occurred, and it must not be read as a 0% rejection rate. The trial also stopped early because of COVID-19 restrictions, with 17 patients analysed, and increased creatinine was among its commonest grade 3-4 adverse events.
  • Barbir 2024 (narrative review) (primary evidence) — While early post-marketing data revealed an allograft rejection risk of 40%-50%, 2 recent small prospective trials have demonstrated lower rates of rejection of 0%-12%, suggesting that maintenance immunosuppression modification prior to ICI start modulates rejection risk.A narrative review's summary bands, with no denominator stated for either. Carried because it is the only source here that states the era trend as a claim rather than leaving it to be inferred from two cohorts a decade apart. Its abstract does not name the two prospective trials, so this card does not claim which they are.

Grade definition

  • CTCAE v5.0 (standard grading vocabulary) — CTCAE v5.0 has no term for allograft rejection. The nearest term is 'Acute kidney injury', which grades the size of the creatinine rise and says nothing about the lesion causing it.This card's four rungs are indexed on that AKI term because the schema requires four rungs, and every rung says so. No management step on this card is indexed to which band a patient falls in — with the single exception of the grade 4 care setting, whose own band text reads 'life-threatening consequences; dialysis indicated'.
  • Legris 2025 (French multicentre kidney cohort) (primary evidence) — All rejection episodes involved T cells, and one-third showed additional humoral features.The transplant literature classifies this lesion by mechanism on biopsy, not by a severity grade. That axis and CTCAE's creatinine bands answer different questions, and only the biopsy axis changes what a transplant team does.

ICI action

  • Carroll 2022 (phase 1 trial, baseline immunosuppression unchanged) (primary evidence) — Maintaining baseline immunosuppression before treatment with an immune checkpoint inhibitor in kidney transplant recipients might not affect expected efficacy and might reduce the risk of allograft rejection mediated by immune checkpoint inhibitors.One prospective trial's conclusion: leave the maintenance regimen alone and the ICI can be given. It is a single-arm phase 1 study of 17 patients that stopped early, and its stated conclusion is hedged twice ('might', 'might').
  • Schenk 2024 (prospective trial, standardized tacrolimus + prednisone) (primary evidence) — In most KTR with advanced skin cancer, TACRO + PRED provides insufficient allograft protection and compromises immune-mediated tumor regression after administration of NIVO ± IPI.The other prospective trial's conclusion, and the opposite one: a standardized regimen neither protected the grafts nor left the tumour response intact. Two trials, two immunosuppression strategies, opposite verdicts on whether the ICI can be given safely — which is why nothing on this ladder resolves them into a rule.
  • Abdel-Wahab 2019 (institutional series and systematic review) (primary evidence) — Overall, 31% of patients permanently discontinued CPIs because of allograft rejection.What treating teams actually did in a series where rejection was the stated reason. It is neither a recommendation nor grade-indexed — the abstract states no rule — and it is shown beside the two trial conclusions rather than resolved with them.

Grade ladder

Grade 1
Hold — consider permanent discontinuationConsider admission

Grade 1 — creatinine 1.5-2.0x above baseline (CTCAE v5.0, Acute kidney injury). The band describes the creatinine rise in the allograft and nothing else: CTCAE has no term for allograft rejection, and no source cited here grades this lesion or indexes any management step to a grade.

Dose not carried

No source cited on this card states a corticosteroid dose for ICI-associated kidney allograft rejection, so no dose and no taper is carried at any rung. The two kidney-scoped doses this repo holds are recommendations about interstitial nephritis, not about alloimmune rejection of a transplanted organ: ASCO's is scoped to grade 2 or higher ICI-related nephritis/AKI, and ASON's ~1 mg/kg/day to clinically suspected ICI-AKI — and although ASON's stated scope does name kidney transplant recipients as a special population, no chapter of it has been read into this atlas, and D-STEROID bars borrowing a dose across lesions inside the kidney exactly as it bars borrowing across organs. The only treatment these sources describe is an increased dose of steroid in a two-case Tier 4 report, which states no number and in which graft function improved in neither patient. 'Unsourced' rather than 'none': no source says a steroid is not indicated.

  • · Allograft biopsy — the review held here states biopsy is required to confirm the diagnosis, and the creatinine band this rung is named for does not make it.
  • · Transplant nephrology and oncology in the same conversation, including an explicit statement of whether the allograft is still to be preserved.
  • · Review the maintenance regimen and any reduction made before the ICI — the only modifiable factor either large dataset identifies.

Escalate when: A biopsy showing rejection, or a creatinine that keeps rising. No source cited here states an escalation trigger for this lesion, and crossing a CTCAE band is not itself a treatment decision here, because nothing on this ladder is indexed to one.

Grade 2
Hold — consider permanent discontinuationConsider admission

Grade 2 — creatinine >2.0-3.0x above baseline (CTCAE v5.0, Acute kidney injury). Same caveat as grade 1: the band is a creatinine descriptor, not a severity of rejection. Nothing on this rung differs from the one below it because no source cited here differentiates them.

Dose not carried

No source cited on this card states a corticosteroid dose for ICI-associated kidney allograft rejection, so no dose and no taper is carried at any rung. The two kidney-scoped doses this repo holds are recommendations about interstitial nephritis, not about alloimmune rejection of a transplanted organ: ASCO's is scoped to grade 2 or higher ICI-related nephritis/AKI, and ASON's ~1 mg/kg/day to clinically suspected ICI-AKI — and although ASON's stated scope does name kidney transplant recipients as a special population, no chapter of it has been read into this atlas, and D-STEROID bars borrowing a dose across lesions inside the kidney exactly as it bars borrowing across organs. The only treatment these sources describe is an increased dose of steroid in a two-case Tier 4 report, which states no number and in which graft function improved in neither patient. 'Unsourced' rather than 'none': no source says a steroid is not indicated.

  • · Allograft biopsy — the review held here states biopsy is required to confirm the diagnosis, and the creatinine band this rung is named for does not make it.
  • · Transplant nephrology and oncology in the same conversation, including an explicit statement of whether the allograft is still to be preserved.
  • · Review the maintenance regimen and any reduction made before the ICI — the only modifiable factor either large dataset identifies.

Escalate when: Graft function not improving after whatever rejection therapy the treating team gives. No source cited here states a trigger for this lesion; this wording is the atlas's non-committal encoding of that silence, not a sourced rule.

Grade 3
Hold — consider permanent discontinuationConsider admission

Grade 3 — creatinine >3.0x baseline or >4.0 mg/dL (CTCAE v5.0, Acute kidney injury). Same caveat. The only thing that changes at this rung is the supportive planning a dialysis-range creatinine forces, not any graded recommendation.

Dose not carried

No source cited on this card states a corticosteroid dose for ICI-associated kidney allograft rejection, so no dose and no taper is carried at any rung. The two kidney-scoped doses this repo holds are recommendations about interstitial nephritis, not about alloimmune rejection of a transplanted organ: ASCO's is scoped to grade 2 or higher ICI-related nephritis/AKI, and ASON's ~1 mg/kg/day to clinically suspected ICI-AKI — and although ASON's stated scope does name kidney transplant recipients as a special population, no chapter of it has been read into this atlas, and D-STEROID bars borrowing a dose across lesions inside the kidney exactly as it bars borrowing across organs. The only treatment these sources describe is an increased dose of steroid in a two-case Tier 4 report, which states no number and in which graft function improved in neither patient. 'Unsourced' rather than 'none': no source says a steroid is not indicated.

  • · Allograft biopsy — the review held here states biopsy is required to confirm the diagnosis, and the creatinine band this rung is named for does not make it.
  • · Transplant nephrology and oncology in the same conversation, including an explicit statement of whether the allograft is still to be preserved.
  • · Review the maintenance regimen and any reduction made before the ICI — the only modifiable factor either large dataset identifies.
  • · Plan for renal replacement therapy if graft function does not recover — in the two-case report an increased steroid dose did not restore it in either patient.

Escalate when: Graft function not improving, or renal replacement therapy becoming necessary. No source cited here states a trigger for this lesion; this wording is the atlas's, not a sourced rule.

Grade 4
Hold — consider permanent discontinuationAdmit

Grade 4 — life-threatening consequences, dialysis indicated (CTCAE v5.0, Acute kidney injury). Same caveat: the band still describes the creatinine course, not a severity of rejection. Both patients in the case report cited here reached this band and neither recovered graft function.

Dose not carried

No source cited on this card states a corticosteroid dose for ICI-associated kidney allograft rejection, so no dose and no taper is carried at any rung. The two kidney-scoped doses this repo holds are recommendations about interstitial nephritis, not about alloimmune rejection of a transplanted organ: ASCO's is scoped to grade 2 or higher ICI-related nephritis/AKI, and ASON's ~1 mg/kg/day to clinically suspected ICI-AKI — and although ASON's stated scope does name kidney transplant recipients as a special population, no chapter of it has been read into this atlas, and D-STEROID bars borrowing a dose across lesions inside the kidney exactly as it bars borrowing across organs. The only treatment these sources describe is an increased dose of steroid in a two-case Tier 4 report, which states no number and in which graft function improved in neither patient. 'Unsourced' rather than 'none': no source says a steroid is not indicated.

  • · Allograft biopsy — the review held here states biopsy is required to confirm the diagnosis, and the creatinine band this rung is named for does not make it.
  • · Transplant nephrology and oncology in the same conversation, including an explicit statement of whether the allograft is still to be preserved.
  • · Review the maintenance regimen and any reduction made before the ICI — the only modifiable factor either large dataset identifies.
  • · Renal replacement therapy where indicated, and an explicit decision about the graft: one reported case required a delayed graft nephrectomy and the other remained dialysis-dependent.

Escalate when: Already the top rung, and no source cited here states an escalation trigger beyond it. Both reported cases reached this band — one required a delayed graft nephrectomy after perioperative bleeding and stabilized on haemodialysis, the other remained dialysis-dependent and died.

Rechallenge

Individualize the decision

No source cited here reports the outcome of re-exposing a kidney transplant recipient to an ICI after ICI-associated allograft rejection, so no recurrence figure is carried and none is borrowed from this atlas's AIN card, which measures a different lesion in a native kidney. The two-case report cited here records that one recipient, after a delayed graft nephrectomy and stabilization on haemodialysis, showed a durable response to the ICI, while the other remained dialysis-dependent and died — but its abstract does not state whether ICI dosing was continued in either patient, so this card does not present that as a rechallenge outcome. Where the graft is still functioning, the maintenance regimen is the only modifiable factor either large dataset identifies. The stance recorded here is this atlas's reading of an evidence base with no re-exposure data, not a recommendation any cited source makes.

  • · An explicit decision, taken with the transplant team, about whether the allograft is still to be preserved — the answer changes what a rechallenge costs.
  • · Allograft biopsy to establish what the first event actually was; the review held here states biopsy is required to confirm rejection.
  • · Reconsideration of the maintenance regimen — mTOR inhibitor use carried a lower rejection risk in the 69-patient kidney cohort (odds ratio 0.26, 95% CI 0.09-0.72), and steroid plus mTOR-inhibitor maintenance carried a lower risk in the 343-recipient individual-participant meta-analysis (hazard ratio 0.30, 95% CI 0.14-0.63).
  • · Serial donor-derived cell-free DNA where available, which preceded the creatinine rise in 2 of the 3 trial patients with treatment-related allograft loss.

Deliberately not carried

  • Linked society recommendations — Left empty. ASCO, ESMO, SITC and ASON are all in this atlas's guideline library, but none of them can be linked here. ASON's stored scope does explicitly name kidney transplant recipients as a special population — so it is NOT true that no society document reachable here touches ICI use in transplant recipients — yet the recommendation this atlas stores from ASON is scoped to ICI-AKI/interstitial nephritis, no ASON chapter has been read into IRAE_GUIDELINE_ORGANS, and check:data 19.5 would reject the link. ASCO's, ESMO's and SITC's stored recommendations are likewise about ICI-related nephritis/AKI. Under D-STEROID that scope does not reach an alloimmune rejection of a transplanted organ, so no society document held here grades this lesion or states a corticosteroid dose for it. The management content on this card is observational, and the grade ladder carries no dose at any rung.
  • Frequency band — Set to 'unclear' rather than a band. Every other entry in this file bands against the ICI-TREATED population; the rates on this card are measured in kidney transplant recipients receiving ICIs, who are a small fraction of that population. A one-word chip cannot carry the change of denominator, and 'Common' beside 25 cards where the word means something else would read as a claim about ICI-treated patients that no source makes. The incidence rows carry the subgroup rates with their denominators instead.
  • Difference between ICI classes — No source cited here reports a class-stratified rejection rate. The 39-patient institutional series and systematic review reports that rejection occurred at similar rates for anti-CTLA-4 and anti-PD-1 therapy, and the 83-case systematic review that outcomes were similar across immunotherapy regimens — but neither prints the stratified numbers behind those statements. A stated absence of difference without the strata is not a measured comparison, so no skew framing is entered.
  • Grade 3 or higher share — No source cited here grades this lesion at all. CTCAE v5.0 has no term for allograft rejection, no society document held here grades it, and none of the cohorts reports a grade 3+ share. A figure would have to be manufactured from the graft-loss proportions, which measure a different thing.
  • Mortality — Null on purpose. The 39-patient institutional series and systematic review prints a death share beside its rejection rate, but its abstract never identifies which population that share is taken over, so it is carried nowhere on this card and no population is supplied for it. The two mortality figures here that do carry denominators measure cancer death rather than death from rejection: the individual-participant meta-analysis reports the share dying of cancers within three years, and the 83-case systematic review states that most of the deaths were related to cancer progression. No source cited here reports a case-fatality rate for the rejection episode.
  • Median time to onset — All three cohorts that report a median report it in DAYS, and they disagree: 24 days in the 69-patient kidney cohort, 52 days in the 34-patient French kidney cohort, 21 days in the 39-patient institutional series and systematic review. Converting one of them to weeks would both change the unit the source used and promote one cohort over the other two, so the slot stays null and the onset note carries all three with their cohort sizes.
  • Onset range — No cited source reports an interquartile range or a full range for time to rejection — only point medians, plus two individual cases at 6 days and 2 weeks. A range is not inferred from that case pair.
  • Onset window — The 1-6 week window is the MODAL value of the three reported medians (21 and 24 days fall inside it), not a consensus: the French cohort's 52 days is about 7.4 weeks and falls in the next bucket up, and the other datasets cited here report no median at all. The note says so explicitly rather than letting the chip imply agreement.
  • Grade definitions on the ladder — The bands are CTCAE v5.0 'Acute kidney injury' as named, and they describe the size of the creatinine rise in the allograft, not a severity of rejection. CTCAE has no term for allograft rejection and no source cited here grades this lesion, so nothing on this ladder is indexed to the band a patient falls in, with the single exception noted in the care-setting row. The four rungs exist because the schema requires four, and each rung says so in its own definition.
  • Corticosteroid — No corticosteroid dose and no taper at any rung. No source cited here states one for this lesion. The kidney-scoped doses this repo holds are ASCO's, scoped to grade 2 or higher ICI-related nephritis/AKI, and ASON's ~1 mg/kg/day, scoped to clinically suspected ICI-AKI; ASON's stated scope does name kidney transplant recipients as a special population, but no chapter of it has been read into this atlas and its recommendation is still about interstitial nephritis. D-STEROID bars borrowing a dose across lesions inside the kidney exactly as it bars borrowing across organs. The only treatment description in these sources is an increased dose of steroid in a two-case Tier 4 report, which states no number and in which graft function improved in neither patient. The rungs therefore carry kind: 'unsourced' rather than kind: 'none', which would assert that no steroid is indicated.
  • ICI action — No source cited here states a grade-indexed rule for holding, resuming or permanently stopping the ICI in this lesion. The value is identical at all four rungs precisely so that no gradient is implied, and the stored value is this atlas's non-committal encoding of an unresolved decision, not a recommendation any source makes. What the sources actually contain is a genuine three-way divergence — two prospective trials reaching opposite conclusions about whether an immunosuppression strategy makes the ICI givable, and one series reporting how often teams stopped it for rejection — shown side by side rather than resolved into a rule.
  • Care setting — No source cited here states a care setting at any grade, so every value in this column is this atlas's encoding rather than a recommendation. The enum has no unsourced state, which is why a value at every rung is itself a pick. The pick is deliberately conservative and deliberately flat: consider-admission at grades 1-3 and admit at grade 4. It does NOT track CTCAE's own disposition clauses (grade 3 reads "hospitalization indicated", grade 4 "dialysis indicated") — a CTCAE-faithful column would read outpatient at the lower bands, as this organ's glomerular card's does. Here the three cohort medians put rejection within weeks of the first dose, graft loss is frequent in every dataset cited, and the biopsy every source requires is rarely arranged from home — so the column leans toward admission at every band. It should not be read as rejection-specific evidence: no source grades this lesion.
  • Second-line therapy — Empty at every rung. No source cited here names any second-line agent for ICI-associated kidney allograft rejection — a term scan of all nine abstracts finds no anti-thymocyte globulin, IVIG, plasmapheresis or rituximab. The only treatment reported is an increased dose of steroid in a two-case report, which did not restore graft function in either patient; the review held here says it covers approaches to treating rejection but its abstract names none, and this card carries only what its sources print.
  • What triggers escalation — No source cited here states an escalation trigger for this lesion at any grade. The four triggers are this atlas's own wording — a biopsy result, a creatinine that keeps rising, graft function that does not recover, dialysis becoming necessary — and each rung says so inline. They are not indexed to a CTCAE band and no cited abstract supplies a threshold for any of them.
  • Whether to rechallenge — 'Individualize the decision' is this atlas's reading, not a sourced recommendation. No source cited here states a rechallenge position for ICI-associated kidney allograft rejection, and none reports re-exposing a recipient after one. The rationale says so in the same words rather than letting the rendered headline imply a recommendation exists.
  • Recurrence on rechallenge — No source cited here reports re-exposure to an ICI after ICI-associated allograft rejection, so no recurrence figure exists to carry. It is deliberately not inferred from the recurrence rate on this atlas's AIN card: that measures a different lesion in a native kidney.
  • Kidney injury signatures — Not set. None of the 14 members of this repo's kidney-injury vocabulary describes alloimmune rejection of a transplanted kidney. Mapping it to 'ain', the nearest interstitial lesion, would file this entry on /injuries/ain as a form of checkpoint interstitial nephritis — which is exactly the confusion the first row of this card's differential exists to prevent. In the TypeScript literal the field is OMITTED entirely rather than set to an empty array.
  • Overlapping syndromes — Empty. The only cluster in IRAE_OVERLAPS is the myocarditis/myositis/myasthenia triple-M syndrome, which does not include this entity, and no source cited here reports a cross-organ syndrome that this lesion belongs to.
  • Links to other organ cards — Empty. This lesion is confined to the transplanted kidney. Tagging it onto another organ's card would put transplant-specific management in front of readers whose sources do not cover transplantation at all.
  • Differential — transplant-specific causes — The transplant-specific differential a nephrologist would actually work through — BK polyomavirus nephropathy, calcineurin-inhibitor toxicity, recurrent glomerular disease, obstruction of the graft — is standard transplant nephrology and is deliberately NOT listed here, because none of the sources cited on this card enumerates it and this atlas does not carry unsourced clinical lists. The three rows that are present are the ones its own sources support.
  • Reported frequency — reconciling the rates — The twelve rows are not averaged or reconciled. They differ by era (2010-2020 versus 2015-2024), by numerator definition (clinically diagnosed rejection, biopsy-proven rejection, graft loss conditional on rejection, graft loss over a whole cohort, treatment-related allograft loss, and one trial's composite endpoint), by organ mix (kidney-only versus solid-organ-wide), by time indexing (crude proportion versus one-year incidence), and by whether maintenance immunosuppression was reduced before the ICI — the variable the two prospective trials were built to isolate and that no retrospective cohort here can. Showing them side by side is the finding, and the rejection rows and the graft-loss rows are kept separate so no reader can read one as the other.

Sources

  • Murakami N, Mulvaney P, Danesh M, Abudayyeh A, Diab A, Abdel-Wahab N, et al. (2021) A multi-center study on safety and efficacy of immune checkpoint inhibitors in cancer patients with kidney transplant.Cited for: The core kidney-specific rejection rate, the timing, the graft-loss fraction among those who rejected, and the two maintenance-immunosuppression associations. The span also carries the design of the non-ICI comparator, because that comparator was assembled stage-matched for the study's efficacy question rather than matched for rejection risk — which is why this card does not turn the 5.4% into a ratio. Note what the abstract does NOT say: 19 of the 29 who rejected lost the allograft, which is two thirds and not 'about half', and the median is 24 days and not 'three weeks'.Supporting text: the PubMed abstract (checkable at the link above)
  • Saleem N, Wang J, Rejuso A, Teixeira-Pinto A, Stephens JH, Wilson A, et al. (2025) Outcomes of Solid Organ Transplant Recipients With Advanced Cancers Receiving Immune Checkpoint Inhibitors: A Systematic Review and Individual Participant Data Meta-Analysis.Cited for: The pooled one-year rejection and graft-loss incidences with a stated denominator, the maintenance-regimen association, and the fact that cancer — not rejection — is what these patients die of. The population is SOLID-ORGAN-wide (70.9% kidney), which is why every row citing it says so.Supporting text: the PubMed abstract (checkable at the link above)
  • Carroll RP, Boyer M, Gebski V, Hockley B, Johnston JK, Kireta S, et al. (2022) Immune checkpoint inhibitors in kidney transplant recipients: a multicentre, single-arm, phase 1 study.Cited for: The one trial that left baseline immunosuppression unchanged, and one half of this card's iciAction divergence. The span is a single ordered excerpt so that four things travel together and none of them can be read without the others: the composite definition of the primary endpoint, the endpoint result, the early stop for COVID-19, and the grade 3-4 adverse-event list in which increased creatinine appears in three of the 17. It also carries the eligibility criterion that limits the trial to recipients with no or low donor-specific HLA antibodies, which is what the pre-dose HLA work-up row rests on.Supporting text: the PubMed abstract (checkable at the link above)
  • Schenk KM, Deutsch JS, Chandra S, Davar D, Eroglu Z, Khushalani NI, et al. (2024) Nivolumab + Tacrolimus + Prednisone ± Ipilimumab for Kidney Transplant Recipients With Advanced Cutaneous Cancers.Cited for: The counterweight to the trial above and the other half of this card's iciAction divergence — standardized tacrolimus plus low-dose prednisone did NOT protect the grafts and compromised tumour regression. The single ordered span keeps the standardized regimen, the dd-cfDNA monitoring interval, the TRAL counts and the dd-cfDNA lead time in one place, so the lead time cannot be read without the interval that makes it actionable. Note that the prednisone 5 mg here is trial maintenance immunosuppression, not treatment of a rejection episode, and cannot set a ladder dose.Supporting text: the PubMed abstract (checkable at the link above)
  • Legris T, Sallée M, Charmetant X, Thaunat O, Matignon M, Joher N, et al. (2025) Immune Checkpoint Inhibitors in Kidney Transplant Recipients: A French Multicenter Retrospective Cohort Study.Cited for: The contemporary (2015-2024) kidney-only rate on a biopsy-proven numerator, the later median onset, the mechanism split on biopsy, and the absence of any identified clinical predictor. Also the finding on this card that cuts hardest against its own headline, and the reason it appears in a rendered field rather than only in a stored span: this cohort reports a LOWER incidence of graft loss (15%) than previously reported rates and concludes that ICI therapy is a viable option for these recipients.Supporting text: the PubMed abstract (checkable at the link above)
  • Abdel-Wahab N, Safa H, Abudayyeh A, Johnson DH, Trinh VA, Zobniw CM, et al. (2019) Checkpoint inhibitor therapy for cancer in solid organ transplantation recipients: an institutional experience and a systematic review of the literature.Cited for: The renal stratum printed with its own denominator (11 of 23), the earliest of the three medians, the absence of a class difference between anti-CTLA-4 and anti-PD-1, the permanent-discontinuation share, and the biopsy findings. This study is an MD Anderson chart review AND a systematic review of five databases whose results are never separated, so no result here is attributed to either half alone. Its abstract also prints a graft-loss share and a death share with no stated population; neither is carried anywhere on this card, and the mortality omission row says why.Supporting text: the PubMed abstract (checkable at the link above)
  • d'Izarny-Gargas T, Durrbach A, Zaidan M (2020) Efficacy and tolerance of immune checkpoint inhibitors in transplant patients with cancer: A systematic review.Cited for: The published-case rejection rate with its denominator, the share of rejections that ended in end-stage organ failure, the framing of the oncology/transplant trade-off that the multidisciplinary work-up row and the ladder's supportive bullets rest on, and — the reason this card carries mortality: null — that most of the deaths were related to cancer progression rather than to rejection.Supporting text: the PubMed abstract (checkable at the link above)
  • Barbir EB, Abdulmoneim S, Dudek AZ, Kukla A (2024) Immune Checkpoint Inhibitor Therapy for Kidney Transplant Recipients - A Review of Potential Complications and Management Strategies.Cited for: The era framing and the biopsy requirement. Its 40%-50% and 0%-12% are a review's summary BANDS with no denominators — carried at review weight, exactly as the AIN card carries an undenominated 2-5% beside a pooled figure — and are not measured rates. Also the source for placing non-invasive biomarkers at pre-ICI risk stratification and at monitoring rather than at diagnosis. It is a narrative review, not a guideline, and it states no corticosteroid dose.Supporting text: the PubMed abstract (checkable at the link above)
  • Morrison SA, Vinson AJ (2024) Acute Allograft Rejection in Kidney Transplant Recipients Treated With Immune Checkpoint Inhibitors: An Educational Case Report.Cited for: EXISTENCE AND PHENOTYPE ONLY — two cases, so no rate and no dose. That rejection can present as AKI 6 days and 2 weeks after ICI initiation in the setting of reduced immunosuppression; that an increased steroid dose restored graft function in neither; and the two endpoints after graft loss, one recipient stabilizing on haemodialysis with a durable response to the ICI and the other remaining dialysis-dependent and dying. The abstract does NOT state whether ICI dosing continued in either patient, and nothing on this card claims it did. The phrase 'an increased dose of steroid' states no number and cannot license a ladder rung under D-STEROID.Supporting text: the PubMed abstract (checkable at the link above)

ICI-associated glomerular disease

Also called: ICI-associated glomerulonephritis · ICI-GN · checkpoint inhibitor glomerular disease · ICI-associated podocytopathy · ICI-related nephrotic syndrome

The kidney irAE a biopsy exists to find. Interstitial nephritis is what ICI-AKI usually is, but a minority of cases are glomerular — pauci-immune glomerulonephritis, podocytopathies and C3 glomerulopathy — and they present with proteinuria rather than a bland sediment, carry worse kidney outcomes, and in 41% of reported cases sit alongside interstitial nephritis rather than instead of it.

RareSeriousonset: any time on therapy

Reported frequency

  • pauci-immune GN and renal vasculitis 27%, podocytopathies 24%, C3 glomerulopathy 11% — lesion mix among biopsy-confirmed ICI-associated glomerular disease

    of 45 biopsy-confirmed cases from 27 published articles · Systematic review and meta-analysis of all biopsy-proven published cases and series, MEDLINE/EMBASE/Cochrane from inception to February 2020

    pooled meta-analysis · 33426386

  • pauci-immune GN 28.3%, podocytopathies 26.4%, immune-complex GN 18.9% — lesion mix in a second published-case pool

    of 53 patients with ICI-related glomerular disease retrieved from PubMed · The rank order matches the pooled review above, but the two are NOT independent samples — both are assembled from published case reports indexed in the same databases over overlapping periods, so the agreement reflects shared sourcing rather than replication

    case series · 36420664

  • 41% — concomitant acute interstitial nephritis

    of 45 biopsy-confirmed cases · Same pooled review. Finding interstitial nephritis on the biopsy does not exclude a glomerular lesion, and finding a glomerular lesion does not exclude interstitial nephritis

    pooled meta-analysis · 33426386

  • acute tubulointerstitial nephritis in 14 of 16 — share of ICI-treated kidney biopsies showing interstitial rather than glomerular disease

    of 16 patients biopsied after checkpoint-inhibitor therapy, 2008-2018 · Single-centre biopsy series. The base rate that makes this entry a second-line consideration rather than a first one — and in that series ATIN frequently CO-OCCURRED with a glomerular lesion rather than excluding it

    case series · 30612580

  • every IgA nephropathy, minimal change disease and crescentic GN case was ICI-related; all crescentic GN was ANCA-negative — attribution of specific glomerular lesions to ICIs in a national biopsy registry

    of 135 anticancer-drug-related cases confirmed from 449 'drug-induced' biopsies across 49 facilities · Japan Renal Biopsy Registry, 2018-2021. Tubulointerstitial nephritis (47.4%) and TMA (35.6%) were the commonest diagnoses overall; the glomerular lesions named here were attributable to checkpoint inhibitors specifically

    cohort · 39664993

Severity and class

  • approximately one-third — death among reported biopsy-confirmed cases

    of 45 biopsy-confirmed cases · Pooled review. These are patients with advanced cancer receiving checkpoint blockade, so this is all-cause mortality in that population and NOT a case-fatality rate for the kidney lesion

    pooled meta-analysis · 33426386

  • 84.1% anti-PD-1 (58.9%) or anti-PD-L1 (25.2%) — share of FAERS ICI-related nephrotic-syndrome reports by checkpoint target

    of 404 FAERS reports of ICI-related nephrotic syndrome · FAERS, Q1 2011 to Q1 2025. A share of REPORTS, not of treated patients, and it tracks how widely each class is prescribed as much as any difference in risk

    pharmacovigilance · 41087308

Onset

No source read here reports an onset distribution for the glomerular lesion specifically. The single-centre biopsy series gives a median 14 weeks (range 6-56) from starting a checkpoint inhibitor to AKI, but 14 of its 16 biopsies were interstitial nephritis, so that is an ICI-AKI figure rather than a glomerular one. The FAERS analysis reports median onset 22.5 days for severe versus 96.5 days for non-severe nephrotic-syndrome REPORTS (p=0.0479) — a comparison between report subgroups, not a description of when the lesion appears. Neither is converted into a window here.

Presentation

  • · Proteinuria is the cardinal abnormality and the axis treatment response is measured on — complete remission of proteinuria in 45% and partial remission in 38% of the 45 pooled biopsy-confirmed cases
  • · Nephrotic syndrome at the podocytopathy end of the spectrum; 404 reports of ICI-related nephrotic syndrome accumulated in FAERS between Q1 2011 and Q1 2025
  • · Acute kidney injury severe enough to require renal replacement therapy in 25% of the pooled cases
  • · Frequently NOT an isolated glomerular lesion: concomitant acute interstitial nephritis in 41% of pooled cases, and in the single-centre series interstitial nephritis appeared alongside pauci-immune GN, membranous GN, C3 GN, IgA nephropathy or AA amyloidosis

Differential

  • · ICI-associated acute tubulointerstitial nephritis, which is far commoner and is what an ICI-AKI is until a biopsy says otherwise — 14 of 16 biopsies in the single-centre series, and tubulointerstitial nephritis was the commonest diagnosis (47.4%) in the national registry
  • · Not an either/or: concomitant AIN was reported in 41% of pooled glomerular cases, so an interstitial infiltrate on the biopsy neither excludes a glomerular lesion nor settles the treatment
  • · Glomerular disease unrelated to the checkpoint inhibitor — the malignancy itself, a second nephrotoxic exposure, or pre-existing kidney disease. The whole evidence base here is published biopsy-confirmed cases, a design that establishes the lesions occur and cannot establish attribution in an individual patient
  • · ANCA-associated vasculitis: the crescentic glomerulonephritis in the national registry was ICI-related and ANCA-NEGATIVE in every case, so a negative ANCA does not argue against a pauci-immune lesion here

Work-up

  • Quantified proteinuria (urine protein-to-creatinine ratio or timed collection) · All patients

    The lesion's cardinal abnormality and the axis response is measured on. Without a baseline, the 45%/38% complete and partial remission figures have nothing to be measured against.

    33426386

  • Serum creatinine trend against baseline · All patients

    Grades the AKI component, which is the only axis the ladder below is written in. It does not capture the proteinuria that defines this lesion — a patient can carry nephrotic-range proteinuria with a stable creatinine and sit below grade 1 entirely.

  • Urinalysis with microscopy · All patients

    Separates the proteinuric, sometimes active sediment of a glomerular lesion from the subnephrotic proteinuria and pyuria typical of interstitial nephritis. A sediment atypical for AIN is part of what moves the decision toward biopsy.

    33426386

  • Kidney biopsy · If atypical

    The only way to separate a glomerular lesion from interstitial nephritis, and every case in this entity's evidence base is biopsy-confirmed. The pooled review's own instruction is to consider obtaining a specimen when features atypical for AIN are present — a clinical phrase, not a numeric grade, which is why this row is indexed to atypia rather than to a CTCAE band. The single-centre series goes further and recommends biopsy for ICI-related AKI generally, in the absence of any predictive blood or urine biomarker.

    33426386, 30612580

  • Nephrology referral · All patients

    The pooled review directs oncologists and nephrologists jointly to be aware of these lesions and to consider biopsy; the treatment decisions below (biopsy timing, immunosuppression beyond corticosteroids, rituximab) are nephrology decisions.

    33426386

Sources differ here

Second-line therapy

  • Alasadi 2026 — single-centre series arm (primary evidence) — In a single-centre series of 16 patients with biopsy-proven non-vasculitic ICI-induced glomerulonephritis, 8 treated with rituximab had a better change in proteinuria response than the 8 treated with corticosteroids alone (P = .029), with a shorter median corticosteroid course (2.5 vs 8 weeks).The rituximab group also started with higher median proteinuria (grade 3 vs 1) and the groups were not randomized, so the comparison is not a clean estimate of effect.
  • Alasadi 2026 — literature-review arm (primary evidence) — In the same paper's literature review of 42 published patients, renal response did not differ significantly between the 10 who received rituximab and the 32 who did not.Same paper, opposite direction. The authors call rituximab an attractive option and state that clinical trials are needed; this card carries both results rather than reconciling them.

Grade ladder

Grade 1
Hold the ICIOutpatient

Grade 1 — creatinine 1.5-2.0x above baseline (CTCAE v5.0, Acute kidney injury). The AKI term is the only axis any source cited here grades, and it does not grade proteinuria: a patient with nephrotic-range proteinuria and a stable creatinine carries this lesion's cardinal abnormality while sitting below this rung entirely.

Dose not carried

No source read here states a corticosteroid dose, route or taper for an ICI-associated GLOMERULAR lesion at any grade. The pooled review establishes only that 98% of 45 biopsy-confirmed cases received corticosteroids and states no regimen. This atlas's stored ASCO recommendation is scoped to ICI-related nephritis/AKI — the tubulointerstitial entity — so carrying its mg/kg bands here would borrow a dose across LESIONS inside one organ, which is D-STEROID's failure one level down from the one it was written for.

  • · Quantify proteinuria and fix a baseline before treatment starts
  • · Exclude alternative causes of glomerular disease and stop concomitant nephrotoxins
  • · Hold rather than continue: the ICI was discontinued in 88% of pooled cases and 89.4% of a second published-case pool, and no source read here stratifies that decision by CTCAE grade for a glomerular lesion

Escalate when: Rising creatinine, new or worsening proteinuria, or a sediment atypical for interstitial nephritis — each of which argues for a biopsy rather than for empirical escalation of immunosuppression.

Grade 2
Hold the ICIOutpatient

Grade 2 — creatinine >2.0-3.0x above baseline (CTCAE v5.0, Acute kidney injury).

Dose not carried

Same as grade 1. Here the LESION rather than the creatinine band decides the regimen, and no source read here pairs the two.

Second line

  • Rituximab, with or without corticosteroids — Biopsy-proven non-vasculitic ICI-induced glomerulonephritis, particularly where the aim is to limit corticosteroid exposure so checkpoint therapy can resume. In a single-centre series of 16 such patients, 8 received rituximab and 8 corticosteroids alone; median corticosteroid duration was 2.5 weeks in the rituximab group against 8 weeks, the change in proteinuria response favoured rituximab (P = .029), and 75% of the rituximab group were rechallenged with an ICI without proteinuria relapse. Read alongside the divergence recorded on this card: in that paper's own literature review of 42 patients, renal response did NOT differ significantly between the 10 who received rituximab and the 32 who did not.
  • · Kidney biopsy where not already obtained — the second-line evidence above is stratified by lesion, not by creatinine
  • · Stop concomitant nephrotoxins and exclude alternative causes

Escalate when: Failure of proteinuria or creatinine to improve, or progression while on corticosteroids.

Grade 3
Hold — consider permanent discontinuationConsider admission

Grade 3 — creatinine >3.0x baseline or >4.0 mg/dL (CTCAE v5.0, Acute kidney injury).

Dose not carried

Same as grades 1 and 2. No cited source states a dose for a glomerular lesion at any severity.

Second line

  • Rituximab, with or without corticosteroids — As at grade 2. The cited series selected patients by LESION and by proteinuria grade, not by AKI grade, so its placement on this ladder is this atlas's arrangement rather than the paper's threshold.
  • · Renal replacement therapy where indicated — 25% of the pooled biopsy-confirmed cases required it
  • · Nephrology involvement

Escalate when: Dialysis-range function, or failure to respond to first- and second-line immunosuppression.

Grade 4
Discontinue permanentlyAdmit

Grade 4 — life-threatening consequences, dialysis indicated (CTCAE v5.0, Acute kidney injury).

Dose not carried

Same as every rung above. The outcome figures attached to this band come from the pooled review's whole 45-case series rather than from a grade-4 subgroup, and are carried in the epidemiology rows rather than presented as this rung's prognosis.

  • · Renal replacement therapy
  • · Nephrology involvement, with the recognition that 19% of the pooled cases remained dialysis-dependent

Escalate when: No further rung — this is the top of the CTCAE band.

Rechallenge

Individualize the decision

The one series that reports rechallenge systematically did so in patients treated with rituximab: 75% of that group were rechallenged with an ICI and had no proteinuria relapse. A second published-case pool records that rechallenge may be considered for cancer progression or as salvage therapy and suggests rechallenging with steroids may be beneficial. Both are small and neither is a controlled comparison, so this is a decision to individualize against the kidney lesion, the degree of recovery and the oncologic alternatives — not a rate to quote at a patient.

  • · The lesion established by biopsy rather than inferred, since the rechallenge evidence is lesion-stratified
  • · Proteinuria response documented before re-exposure — it is the axis on which relapse was measured
  • · Nephrology and oncology agreement on the alternatives, given that the whole evidence base is uncontrolled

Deliberately not carried

  • Reported frequency — No source read here reports an incidence of ICI-associated glomerular disease with a population denominator — that is, a share of ICI-TREATED PATIENTS. Every figure on this card is either a share of biopsy-confirmed published cases or a share of biopsies performed, both of which are conditioned on someone having decided to biopsy. The pooled review's own framing gives 2% to 5% for kidney irAEs as a whole, which is not this lesion. The incidence rows carried above are labelled by their real denominators for that reason.
  • Grade 3 or higher share — No source read here reports a grade 3-or-higher share for this lesion. The pooled review reports renal replacement therapy in 25% and dialysis dependence in 19%, which are outcomes rather than a CTCAE severity distribution, and they are carried as epidemiology rather than converted into one.
  • Median time to onset — See the onset note. The two candidate figures are an ICI-AKI median from a series that was 14/16 interstitial nephritis, and a comparison of median onset between severe and non-severe FAERS reports. Neither describes when a glomerular lesion appears.
  • Corticosteroid — All four rungs. 98% of the pooled biopsy-confirmed cases received corticosteroids and no source read here states the dose, route, taper or duration for a glomerular lesion. The atlas's ASCO recommendation is scoped to ICI nephritis/AKI and does not license a dose for a different lesion in the same organ.
  • ICI action — The hold at grades 1-2 and the permanent discontinuation at grade 4 are this atlas's encoding, not a graded recommendation. What the sources support is that the ICI was stopped in 88% and 89.4% of two published-case pools; neither stratifies that by CTCAE grade.
  • Recurrence on rechallenge — No source read here reports a recurrence RATE with a denominator of rechallenged patients for this lesion. The rituximab series reports that 75% of its 8-patient rituximab group were rechallenged and had no proteinuria relapse — a numerator without a comparable denominator across the whole entity, and drawn from the treated arm of a series selected for rituximab use. It is stated in the rationale rather than promoted to a rate.
  • Overlapping syndromes — renal-ain is the lesion this entry is most often confused with and most often found alongside — concomitant AIN in 41% of pooled cases — but overlapIds addresses the atlas's formal OVERLAP records (triple-M and the like), and no overlap record covers this pair. The relationship is carried in differential[] and presentation[] instead, where a reader will actually meet it.

Sources

  • Kitchlu A, et al. (2021) A Systematic Review of Immune Checkpoint Inhibitor-Associated Glomerular Disease.Cited for: Lesion mix, concomitant-AIN rate, kidney and proteinuria outcomes, and the biopsy trigger for ICI-associated glomerular disease.Supporting text: the PubMed abstract (checkable at the link above)
  • Mamlouk O, et al. (2019) Nephrotoxicity of immune checkpoint inhibitors beyond tubulointerstitial nephritis: single-center experience.Cited for: The base rate of interstitial versus glomerular lesions on ICI-era kidney biopsies, their co-occurrence, and the biopsy recommendation in the absence of a predictive biomarker.Supporting text: the PubMed abstract (checkable at the link above)
  • Alasadi Y, et al. (2026) Rituximab for the treatment of immune checkpoint inhibitor-induced glomerulonephritis.Cited for: Rituximab as a steroid-sparing second line in biopsy-proven non-vasculitic ICI-induced glomerulonephritis, the rechallenge experience, and the divergence between the paper's own two arms.Supporting text: the PubMed abstract (checkable at the link above)
  • Kuwabara T, et al. (2024) SUrvey of renal Biopsy registry database and Anticancer dRUg therapy in Japan (SUBARU-J study).Cited for: Attribution of IgA nephropathy, minimal change disease and crescentic glomerulonephritis to checkpoint inhibitors in a national biopsy registry, and the ANCA-negative character of the crescentic cases.Supporting text: the PubMed abstract (checkable at the link above)
  • Ye J, et al. (2025) Immune checkpoint inhibitor-induced nephrotic syndrome: a pharmacovigilance analysis of 404 FAERS cases and literature case series.Cited for: Volume and checkpoint-target distribution of reported ICI-related nephrotic syndrome, and the onset comparison between severe and non-severe reports.Supporting text: the PubMed abstract (checkable at the link above)
  • He X, et al. (2022) Glomerular diseases after immune checkpoint inhibitors use: What do We know so far?Cited for: A second published-case pool of the lesion mix, the discontinuation and steroid rates, and the rechallenge posture.Supporting text: the PubMed abstract (checkable at the link above)

This atlas covers kidney irAEs in depth

Acute interstitial nephritis

The injury page: every culprit agent, biopsy signature, and management evidence.

ICI nephritis deep dive

The full monograph: mechanism, landmark cohorts, steroids, and rechallenge.

Renal outcomes & rechallenge

Recovery and rechallenge evidence across the checkpoint inhibitors.

Educational use only. Educational synthesis of the published literature — not a treatment protocol, dosing guide, or medical advice. Regimens and agents shown are illustrative of what the literature describes; verify against current guidelines (ASON / KDIGO / ASCO / NCCN) and individualize to the patient. Using this site creates no clinician–patient relationship.

Spotted an error? Tell us. Sourcing and method are described on Methods.